Skip to main content

Full text of "Structure and function in primitive society, essays and addresses;"

See other formats



Essays and Addresses 




With a Foreword by 







Acknowledgment for permission to print the essays and 
addresses in this volume is hereby made to the following : 
The South African Association for the Advancement of 
Science, the Iowa Law Review, the Royal Anthropological 
Institute, the International African Institute, the Fourth 
Pacific Science Congress, the Syndics of the Cambridge 
University Press, the American Anthropological Association 
and the Macmillan Co., New York. 






PROFESSOR Radcliffe-Brown has never had much regard 
for what he calls ' the odd things that I have written 
from time to time ' ; his major interest has been in conveying 
ideas directly to students and colleagues by personal contacts. In 
this he has been eminently successful. He has taught social an- 
thropology at Cambridge, London, Birmingham, Pretoria, 
Johannesburg, Cape Town, Sydney, Yenching, Oxford, Sao 
Paulo, Alexandria and Grahamstown, and in each of these places 
he is remembered with affection and respect. The indebtedness 
of his students has been shown in two collections of essays — one 
American and one English — w-ritten in his honour. And there has 
hardly been a book or article on social anthropology published 
during the last quarter of a century which does not illustrate, 
directly or indirectly, his teaching. 

An examination of the essays in this volume will suggest that 
his writings have been just as influential as his personal contacts. 
He has not, considering that he has been engaged in teaching and 
research in social anthropology for almost fifty years, written 
as much as most persons of his academic eminence. What he has 
written, however, has been faultless. We do not mean by this that 
we necessarily accept his methods or conclusions in detail, but 
rather that the point of view he expresses could not have been 
better expressed. Each of the essays is perfect in conception 
and in expression, and they are tied together by a consistency 
and direction which is rare in modern anthropology. 

We believe that the publication of these essays will be of 
value for several reasons. In the first place, they show the 
development of the thought of a distinguished anthropologist 
for the last twenty-five years, and at the same time illustrate 
some of the more important changes in the orientation of social 
anthropology, with which Professor Radcliffe-Brown was so 
intimately associated during this period. These essays have 
also demonstrated their value in the training of graduate students 
in our major centres for social anthropology. ^The individual 
papers are widely spread in time and place, and frequently 

&74:0 -0 

VI stkuctuul: and i-unction in primitive society 

difHcult to consult. We feci that in presenting this collection 
of essays we are not only showing our esteem for Professor 
RaciclifTc-Rrown, but arc also providing a book which will be 
valualilc to students of social anthropology for a long time to come. 




— -^ Introduction i '^ 

I The Mother's Brother in South Africa 15 ^^ZH 

II Patrilineal and Matrilineal Succession 32 ^35" 

III The Study of Kinship Systems 49 i<^^i 

IV On Joking Relationships 90 i<^^Q 
V A Further Note on Joking Relationships 105 \^M^ 

VI The Sociological Theory of Totemism 117 i1i9 

VII Taboo 133 1^3^ 

VIII Religion and Society 153 i^^-J.r 

IX On the Concept of Function in Social Science 178 i35 

X On Social Structure — 188 ' 

XI Social Sanctions 205 /?^3 

:I Primitive Law 212 i^S3 


THE papers reprinted here are occasional papers in the 
fullest sense of the term ; each of them was written for a 
particular occasion. They do, however, have some 
measure of unity as being written from a particular theoretical 
point of view. 

What is meant by a theory is a scheme of interpretation 
which is applied, or is thought to be applicable, to the under- 
standing of phenomena of a certain class. A theory consists 
of a set of analytical concepts, which should be clearly defined 
in their reference to concrete reality, and which should be logically 
connected. I propose, therefore, by way of introduction to these 
miscellaneous papers, to give definitions of certain concepts of 
which I make use for purposes of analysis of social phenomena. 
It must be remembered that there is very little agreement amongst 
anthropologists in the concepts and terms they use, so that this 
Introduction and the papers that follow are to be taken as an 
exposition of one particular theory, not of a commonly accepted 

History and Theory 

The difference between the historical study of social institutions 
and the theoretical study can be easily seen by comparing economic 
history and theoretical economics, or by comparing the history 
of law with theoretical jurisprudence. In anthropology, however, 
there has been and still is a great deal of confusion which is main- 
tained by discussions in which terms such as 'history' and 'science' 
or 'theory' are used by disputants in very different meanings. 
These confusions could be to a considerable extent avoided by 
using the recognised terms of logic and methodology and dis- 
tinguishing between tdto^raph ic_an. d nomothetic enquiries. 

• In an idiographic enquiry the purpose is to establish as 
acceptable certain particular or factual propositions or statements. 
A nomothetic enquiry, on the contrary, has for its purpose to 
arrive at acceptable general propositions] We define the nature 
of an enquiry by the kind of conclusionstnat are aimed at. 


Ajjstory, as usually understood, is the study of records and 
monuments for the purpose of providing knowledge about con- 
ditions and events of the past, including those investigations that 
are concerned with the quite recent past. It is clear that history 
consists primarily of idiographic enquiriesi In the last century 
there was a dispute, the famous Methodenstreit, as to whether 
historians should admit theoretical considerations in their work 
or deal in generalisations. A great many historians have taken 
the view that nomothetic enquiries should not be included in 
historical studies, which should be confined to telling us what 
happened and how it happened. Theoretical or nomothetic 
enquiries should be left to sociology. But there are some writers 
who think that a historian may, or even should, include theoretical 
interpretations in his account of the past. Controversy on this 
subject, and on the relation between history and sociology, still 
continues after sixty years. Certainly there are writings by his- 
torians which are to be valued not solely as idiographic accounts 
of the facts of the past but as containing theoretical (nomothetic) 
interpretations of those facts. The tradition in French historical 
studies of Fustel de Coulanges and his followers, such as Gustave 
Glotz, illustrates this kind of combination. Some modern writers 
refer J4? it as sociological history or historical sociolog)'. 

|n_anthropology, meaning by that the study of what are 
called the primitive or backward peoples, the term ethnography 
applies to what is specifically a mode of idiographic enquiry, 
the aim of which is-to-^ive acceptable accounts of such peoples 
and their social lifei(Ethnography differs from history in that 
the ethnographer derives his knowledge, or some major part of 
it, from direct observation of or contact with the people about 
whom he writes, and not, like the historian, from written records.] 
Prehistoric archaeology, which is another branch of anthropology, 
is clearly an idiographic study, aimed at giving us factual know- 
ledge about the prehistoric past. 

The theoretical study of social institutions in general is 
usually referred to as sociology, but as this name can be loosely 
used for many different kinds of writings about society we can 
speak more specifically of theoretical or comparative sociology. 
When Frazer gave his Inaugural Lecture as the first Professor 
of Social Anthropology in 1908 he defined social anthropology 
as that branch of sociology that deals with primitive societies. 


Certain confusions amongst anthropologists result from the 
failure to distinguish between historical explanation of institutions 
and theoretical understanding. If we ask why it is that a certain 
institution exists in a particular society the appropriate answer is 
a historical statement as to its origin. To explain why the United 
States has a political constitution with a President, two Houses 
of Congress, a Cabinet, a Supreme Court, we refer to the history 
of North America. This is historical explanation in the proper 
sense of the term. The existence of an institution is explained by 
reference to a complex sequence of events forming a causal chain 
of which it is a result. 

The acceptability of a historical explanation depends on the 
fullness and reliability of the historical record. In the primitive 
societies that are studied by social anthropology there are no 
historical records. We have no knowledge of the development 
of social institutions among the Australian aborigines for example. 
Anthropologists, thinking of their study as a kind of historical 
study, fall back on conjecture and imagination, and invent 
'pseudo-historical' or 'pseudo-causal' explanations. We have had, 
for example, innumerable and sometimes conflicting pseudo- 
historical accounts of the origin and development of the totemic 
institutions of the Australian aborigines. In the papers of this 
volume mention is made of certain pseudo-historical speculations. 
The view taken here is that such speculations are not merely 
useless but are worse than useless. This does not in any way imply 
the rejection of historical explanation but quite the contrary. 

Comparative sociology, of which social anthropology is a 
branch, is here conceived as a theoretical or nomothetic study 
of which the aim is to provide acceptable generalisations. The 
theoretical understanding of a particular institution is its inter- 
pretation in the light of such generalisations. 

Socia l Proce ss 

A first question that must be asked if we are to formulate 
a systematic theory of comparative sociology is: ^^at is the 
concrete, observable, phenomenal reality with which the theory 
is to be concerned? Some anthropologists would say that the 
reality consists of 'societies' conceived as being in some sense 
or other discrete real entities. Others, however, describe the 
reality that has to be studied as consisting of 'cultures', each of 


which is again conceived as some kind of discrete entity. Still 
others seem to think of the subject as concerned with both 
kinds of entities, 'societies' and 'cultures', so that the relation of 
these then presents a problem. ^^ , , 

My own view^ is that the concrete reality with w^hich the social 
anthropologist is concerned in observation, description, com- 
parison and classification, is not any sort of entity but a process, 
the process of social life. The unit of investigation is the social 
life of some particular region of the earth during a certain period 
of time. The process itself consists of an immense multitude of 
actions and interactions of human beings, acting as individuals or 
in combinations or groups. Amidst the diversity of the particular 
events there are discoverable regularities, so that it is possible 
to give statements or descriptions of certzin general features of the 
social life of a selected region. A statement of such significant 
general features of the process of social life constijtutes a descrip- 
tion of what may be called a fQzm-afsociah4^e.\iMy conception 
of social anthropology is as the comparative theoretical study of 
forms of social life amongst primitive peoples. ; ) 

A form of social life amongst a certain collection of human 
beings may remain approximately the same over a certain period. 
But over a sufficient length of time the form of social life itself 
undergoes change or modification. Therefore, while w^e can regard 
the events of social life as constituting a process, there is over and 
above this the process of change in the form of social life. In a 
synchronic description we give an account of a form of social life 
as it exists at a certain time, abstracting as far as possible from 
changes that may be taking place in its features. A diachronic 
account, on the other hand, is an account of such changes over a 
periodiJ5n comparative sociology we have to deal theoretically 
with the continuity of, and with changes in, forms of social life. 


Anthropologists use the word 'culture' in a number of 
different senses. It seems to me that some of them use it as 
equivalent to what I call a form of social life. In its ordinary use 
in English 'culture', which is much the same idea as cultivation, 
refers to a process, and we can define it as the process by which 
a person acquires, from contact with other persons or from such 
things as books or works of art, knowledge, skill, ideas, beliefs, 


tastes, sentiments. In a particular society we_can-jlisiiOArjerxertain 
processes of cultural tradition, using the word tradition in its 
literal meaning of handing on or handing down. The under- 
standing and use of a language is passed on by a process of cultural 
tradition in this sense. An Englishman learns by such a process 
to understand and use the English language, but in some sections 
of the society he may also learn Latin, or Greek, or French, or 
Welsh. In complex modern societies there are a great number of 
separate cultural traditions. By one a person may learn to be a 
doctor or surgeon, by another he may learn to be an engineer 
or an architect. In the simplest forms of social life the number of 
separate cultural traditions may be reduced to two, one for men 
and the other for women. 

If we treat the social reality that we are investigating as being 
jigt an entity but a process, then culture^and cultural tradition 
"are names for certain recognisable aspects of that process, -but 
not, of course, the whole process. The terms are convenient 
ways of referring to certain aspects of human social life. It is by 
reason of the existence of culture and cultural traditions that 
human social life differs very markedly from the social life of 
other animal species. The transmission of learnt ways of thinking, 
feeling and acting constitutes the cultural process, which is a 
specific feature of human social life. It is, of course, part of that 
process of interaction amongst persons which is here defined 
as the social process thought of as the social reality. Continuity 
and change in the forms of social life being the subjects of in- 
vestigation of comparative sociology, the continuity of cultural 
traditions and changes in those traditions are amongst the things 
that have to be taken into account. 

Social System 

It was Montesc[uieu- who, in the middle of the eighteenth 
century, laid~tKe foundations of comparative sociology, and in 
doing so formulated and used a conception that has been and 
can be referred to by the use of the term social system. His theory, 
which constituted what Comte later called 'the first law of social 
statics', was that in a particular form of social life there are relation^ 
of interconnection and interdependence, or what Comte called 
relations oF soirdai"ity^ amongst the various features. The idea 
of a natural or phenomenal system is that of a set of relations 


amongst events, just as a logical system, such as the geometry 
of Euclid, is a set of relations amongst propositions, or an ethical 
system a set of relations amongst ethical judgments. When one 
speaks of the 'banking system' of Great Britain this refers to the 
fact that there is a considerable number of actions, interactions 
and transactions, sucu for example as payments by means of a 
signed cheque drawn on a bank, which are so connected that 
they constitute in their totality a process of which we can make 
an analytical description which will show how they are inter- 
connected and thus form a system. We are dealing, of course, with 
a process, a complex part of the total social process of social life 
in Great Britain. 

I In these essays I have referred to 'kinship systems'. The 
idea is that in a given society we can isolate conceptually, if 
not in reality, a certain set of actions and interactions amongst 
persons which are determined by the relationships by kinship or 
marriage, and that in a particular society these are interconnected 
in such a way that we can give a general analytical description 
of them as constituting a system. The theoretical significance of 
this idea of systems is that our first step in an attempt to under- 
stand a regular feature of a form of social life, such as the use of 
cheques, or the custom by which a man has to avoid social 
contact with his wife's mother, is to discover its place in the 
system of which it is part?7 

The theory of Montesquieu, however, is what we may call a 
theory of a total social system, according to which all the features 
of social life are united into a coherent whole. As a student of 
jurisprudence Montesquieu was primarily concerned with laws, 
and he sought to show that the laws of a society are connected 
with the political constitution, the economic life, the religion, 
the climate, the size of the population, the manners and customs, 
and what he called the general spirit {esprit general) — what later 
writers have called the 'ethos' of the society. A theoretical law, 
such as this 'fundamental law of social statics', is not the same 
thing as an empirical law, but is a guide to investigation. It 
gives us reason to think that we can advance our understanding 
of human societies if we investigate systematically the inter- 
connections amongst features of social life. 


Statics and Dynamics 

Comte poTnfed out that in sociology, as in other kinds of 
science, there are two sets of problems, which he called problems 
of statics and problems of dynamics. In statics we attempt to 
discover and define conditions of existence or of co-existence; y 
in dynamics we try to discover conditions of change. The con- 
ditions of existence of molecules or of organisms are matters 
of statics, and similarly the conditions of existence of societies, 
social systems, or forms of social life are matters for social 
statics. Whereas the problems of social dynamics deal with 
the conditions of change of forms of social life. 

The basis of science is systematic classification. It is the ] 
first task of social statics to make some attempt to compare j 
forms of social life in order to arrive at classifications. But forms \ 
of social life cannot be classified into species and genera in the way 
we classify forms of organic life; the classification has to be not [ 
specific but typological, and this is a more complicated kind of 1 
investigation. It can only be reached by means of the establishing 
of typologies for features of social life or the complexes of features 
that are given in partial social systems. Not only is the task 
complex but it has been neglected in view of the idea that the 
method of anthropology should be a historical method. 

But though the typological studies are one important part 
of social statics, there is another task, that of formulating general- 
isations about the conditions of existence of social systems, or 
of forms of social life.\The so-called first law of social statics is a 
generalisation affirmingThit for any form of social life to persist or 
continue the various features must exhibit some kind and measure 
of coherence or consistence, but this only defines the problem of 
social statics, which is to investigate the nature of this coherence. 

The study of social dynamics is concerned with establishing 
generalisations about how social systems change. It is a corollary 
of the hypothesis of the systematic connection of features of social 
life that changes in some features are likely to produce changes 
in other features. — - 

Social Evolution 

/The tKefiry of social evolution was formulated by Herbert 
Spencer as part of his formulation of the general theory of 
evolution. According to that theory the development of life on 


the earth constitutes a single process to which Spencer applied 
the term 'evolution'. The theory of organic and super-organic 
(social) evolution can be reduced to two essential propositions: 
(i) Thr.t both in the development of forms of organic life and in 
the development of forms of human social life there has been a 
process of diversification.- by which many different forms of 
organic life oTof social life have been developed out of a very 
much smallej number of original forms. (2) That there has been 
a general trend of development by which more complex forms of 
structure and organisation (organic or social) have arisen from 
simpler forms. The acceptance of the theory of evolution only 
requires the acceptance of these propositions as giving us a scheme 
of interpretation to apply to the study of organic and social life. 
But it must be remembered that some anthropologists reject the 
hypothesis of evolution. We can give provisional acceptance to 
Spencer's fundamental theory, while rejecting the various pseudo- 
historical speculations which he added to it, and that acceptance 
gives us certain concepts which may be useful as analytical tools. 

Ada ptation ^ 

ynpiis is a key concept of the theory of evolution. It is, or 
can be, applied both to the study of the forms of organic life 
and to the forms of social life amongst human beings. K living 
organism exists and continues in existence only if it is both 
internally and externally adapted. The internal adaptation 
depends on the adjustment of the various organs and their 
activities, so that the various physiological processes constitute 
(j a continuing functioning system by which the life of the organism 
is maintained. The external adaptation is that of the organism 
to the environment within which it lives. The distinction of 
external and internal adaptation is merely a way of distinguishing 
two aspects of the adaptational system which is the same for 
organisms of a single species^ 

When we come to the social life of animals another feature 
of adaptation makes its appearance. The existence of a colony 
of bees depends on a combination of the activities of the in- 
dividual worker bees in the collection of honey and pollen, the 
manufacture of wax, the building of the cells, the tending of eggs 
and larvae and the feeding of the latter, tha protection of the 
store of honey from robbers, the ventilation of the hive by fanning 


with their wings, the maintenance of temperature in the winter 
by clustering together. Spencer uses the term 'co- operation' to 
refer to this feature of social life. Soc ial life and social adaptat ion] 
therefo re involve the adjustment of the behaviour of indiv idual] 
or ganisms to the requirements of th e_pri}r<^^« ^^y wfiirh thel 
social life continues. ' 

WEeiPw^rexaTrTtne a form of social life amongst human beings 
as an adaptational system it is useful to distinguish three aspects , 
of the total system. There is the way in which the social life is 
adjusted to the physical environment, and we can, if we wish, 
speak of this as the cecological adaptation. Secondly, there are 
the institutional arrangements by which an orderly social life is 
maintained, so that what Spencer calls co-operation is provided 
for and conflict is restrained or regulated. This we might call, 
if we wished, the institutional aspect of social adaptation. Thirdly, 
there is the social process by which an individual acquires habits 
and mental characteristics that fit him for a place in the social 
life and enable him to participate in its activities. This, if we wish, 
could be called cultural adaptation, in accordance with the 
earlier definition of cultural tradition as process. Wliat must be 
emphasised is that these modes of adaptation are only different 
aspects from which the total adaptational system can be looked 
at for convenience of analysis and comparison. 

The theory of social evolution therefore makes it a part of our 
scheme of interpretation of social systems to examine any given 
system as an adaptationa l system. The stability of the system, 
and therefore its continuance over a certain period, depends oq/' 
the effectiveness of the adaptation. 

Social Structure 

The theory of evolution is one of a trend of development by 
which more complex types of structure come into existence by 
derivation from less complex ones. An address on Social Structure 
is included in this volume, but it was delivered in war time and 
was printed in abbreviated form, so that it is not as clear as it 
might be. When we use the term structure we are referring to some 
soVt of ordered arrangement of parts or components. A musical 
composition has a structure, and so does a sentence. A building 
has a structure, so does a molecule or an animal. The components 
or units of social structure are persons, and a person is a human 



being considered not as an organism but as occupying position 
irya social structure. 

J One of the fundamental theoretical problems of sociology is 
that of the nature of social continuity. (^Continuity in forms 
of social life depends on structural continuity, that is, some 
sort of continuity in the arrangements of persons in relation 
to one another.~^t the present day there is an arrangement of 
persons into nations, and the fact that for seventy years I have 
ijelonged to the English nation, although I have lived much of 
my life in other countries, is a fact of social structure. A nation, 
a tribe, a clan, a body such as the French Academy, or such as the 
Roman Church, can continue in existence as an arrangement of 
persons though the personnel, the units of which each is com- 
posed, changes from time to time. There is continuity of the 
structure, just as a human body, of which the components are 
molecules, preserves a continuity of structure though the actual 
molecules, of which the body consists, are continually changing. 
In the political structure of the United States there must always 
be a President; at one time it is Herbert Hoover, at another time 
Franklin Roosevelt, but the structure as an arrangement remains 

V/The social relationships, of which the continuing network 
constitute social structure, are not haphazard conjunctions of 
individuals, but are determined by the social process, and any 
relationship is one in which the conduct of persons in their inter- 
actions with each other is controlled by norms, rules or patterns. 
So that in any relationship within a social structure a person 
knows that he is expected to behave according to these norms and 
is justified in expecting that other persons should do the same. 
; -The established norms of conduct of a particular form of social 
/|life it is usual to refer to as institutions. An institution is an es- 
vtablished norm of conduct recognised as such by a distinguishable 
^ocial group or class of which therefore it is an institution. The 
institutions refer to a distinguishable type or class of social 
relationships and interactions. Thus in a given locally defined 
society we find that there are accepted rules for the way a man is 
expected to behave towards his wife and children. The relation 
of institutions to social structure is therefore twofold. On the one 
side there is the social structure, such as the family in this instance, 
for the constituent relationships of which the institutions provide 


the norms; on the other there is the group, the local society in 
this instance, in which the norm is established by the general 
recognition of it as defining proper behaviour. Institutions, 
if that term is used to refer to the ordering by society of the inter- 
actions of persons in social relationships, have this double 
connection with structure, with a group or class of which it can 
be said to be an institution, and with those relationships within the 
structural system to which the norms apply. In a social system 
there may be institutions which set up norms of behaviour for a 
king, for judges in the fulfilment of the duties of their office, for 
policemen, for fathers of families, and so on, and also norms of 
behaviour relating to persons who come into casual contact within 
the social life. 

A brief mention may be made of the term organisation. The 
concept is clearly closely related to the concept of social structure, 
but it is desirable not to treat the two terms as synonymous. A 
convenient use, which does not depart from common usage in 
English, is to define social structure as an arrangement of persons 
in institutionally controlled or defined relationships, such as the 
relationship of king and subject, or that of husband and wife, and 
to use organisation as referring to an arrangement of activities. 
The organisation of a factory is the arrangement of the various 
activities of manager, foremen, workmen within the total activity 
of the factory. The structure of a family household of parents, 
children and servants is institutionally controlled. The activities 
of the various members of the persons of the household will 
probably be subject to some regular arrangement, and the or- 
ganisation of the life of the household in this sense may be different 
in different families in the same society. The structure of a modern 
army consists, in the first place, of an arrangement into groups — 
regiments, divisions, army corps, etc., and in the second place an 
arrangement into ranks — generals, colonels, majors, corporals, 
etc. The organisation of the army consists of the arrangement of 
the activities of its personnel whether in time of peace or in time 
of war. Within an organisation each person may be said to have 
a role. Thus we may say that when we are dealing with a structural 
system we are concerned with a system of social positions, while 
in an organisation we deal with a system of roles. 


SocialFunciicm^ — 

The term function has a very great number of different 
meanings in different contexts. In mathematics the word, as 
introduced by Euler in the eighteenth century, refers to an 
expression or symbol which can be written on paper, such as 
'log. x\ and has no relation whatever to the same word as used 
in such a science as physiology. In physiology the concept of 
function is of fundamental importance as enabling us to deal with 
the continuing relation of structure and process in organic life. 
A complex organism, such as a human body, has a structure 
as an arrangement of organs and tissues and fluids. Even an 
organism that consists of a single cell has a structure as an arrange- 
ment of molecules. An organism also has a life, and by this we 
refer to a process, ^he concept of organic function is one that is 
used to refer to the CDnnection between the structure of an organ- 
ism and the life process of that organisnD The processes that go 
on within a human body while it is living are dependent on the 
organic structure. It is the function of the heart to pump blood 
through the body. The organic structure, as a living structure, 
depends for its continued existence on the processes that make up 
the total life processes. If the heart ceases to perform its function 
the life process comes to an end and the structure as a living 
structure also comes to an end. Thus process is dependent on 
structure and continuity of structure is dependent on process. 

In reference to social systems and their theoretical under- 
standing one way of using the concept of function is the same as 
its scientific use in physiology. It can be used to refer to the 
interconnection between the social structure and the process of 
social life. It is this use of the word function that seems to me to 
make it a useful term in comparative sociology !^he three concepts 
of process, structure and function are thus components of a single 
theory as a scheme of interpretation of human social systems. 
The three concepts are logically interconnected, since 'function' 
is used to refer to the relations of process and structure. The theory 
is one that we can apply to the study both of continuity in forms 
of social life and also to processes of change in those forms. 

If we consider such a feature of social life as the punishment 
of crime, or in other words the application, by some organised 
procedure, of penal sanctions for certain kinds of behaviour, and • 


ask what is its social function, we have a fundamental problem 
of comparative sociology towards which a first contribution was 
made by Durkheim in his Division du Travail Social. A very wide 
general problem is posed when we ask what is the social function 
of religion. As it has been pointed out in one of the papers in this 
volume, the study of this problem requires the consideration of a 
large number of more limited problems, such as that of the social 
function of ancestor worship in those societies in which it is found. 
But in these more limited investigations, if the theory here 
outlined is accepted, the procedure Kas to be the examination 
of the connection between the structural features of the social 
life and the corresponding social process as both involved in a 
continuing system. 

The first paper in this collection may serve to illustrate these 
theoretical ideas. It deals with an institution by which a sister's 
son is allowed privileged familiarity in his conduct towards his 
mother's brother. The custom is known in tribes of North America 
such as the Winnebago and others, in peoples of Oceania, such as 
the inhabitants of Fiji and Tonga, and in some tribes of Africa. 
My own observations on this institution were made in Tonga and 
Fiji, but as the paper was addressed to a South African audience 
it seemed preferable to refer to a single South African example, 
since a wider comparative discussion would have called for a much 
longer essay. The usual way of dealing with this institution, both 
in Oceania and in Africa, was to offer a pseudo-historical ex- 
planation to the effect that it was a survival in a patrilineal society 
from a former condition of mother-right. 

The alternative method of dealing with the institution is to look 
for a theoretical understanding of it as a part of a kinship system of 
a certain type, within which it has a discoverable function. We 
do not yet have a systematic general typology of kinship systems, 
for the construction of such is a laborious undertaking. I have 
indicated some partial and provisional results of such an attempt 
to determine types in a recent publication in the form of an 
Introduction to a book on African Systems of Kinship and 
Marriage. Amongst the great diversity of kinship systems we can, 
I think, recognise a type of what we may call father-right, and 
another of mother-right. In both these types the kinship structure 
is based on lineages with maximum emphasis on lineage relation- 
ships. In mother-right the lineage is matrilineal, a child belonging 


to the lineage of the mother. Practically all the jural relations 
of a man are those with his matrilineal lineage and its members, 
and therefore he is largely dependent on his mother's brothers, who 
exercise authority and control over him and to whom he looks for 
protection and for inheritance of property. In a system of father- 
right, on the other hand, a man is largely dependent on his 
patrilineal lineage and therefore on his father and father's brothers, 
who exercise authority and control over him, while it is to them 
that he has to look for protection and for inheritance. Father-right 
is represented by the sysyefn of patria potestas of ancient Rome, 
and there are systems that approximate more or less closely to the 
type to be found in Africa and elsewhere. We may regard the 
BaThonga as so approximating. Mother-right is represented 
by the systems of the Nayar of Malabar and the Menangkubau 
Malays, and again there are systems elsewhere that approximate 
to the type. 

The point of the paper on the mother's brother may be said 
to be to contrast with the explanation by pseudo-history the 
interpretation of the institution to which it refers as having a 
function in a kinship system with a certain type of structure. If 
I were to rewrite the paper after thirty years I should certainly 
modify and expand it. But it has been suggested to me that the 
paper may have a certain minor historical interest in relation to 
the development of thought in anthropology and it is therefore 
reprinted almost as it was written with only minor alterations. 

Any interest this volume may have will probably be as an 
exposition of a theory, in the sense in which the word theory is 
here used as a scheme of interpretation thought to be applicable 
to the understanding of a class of phenomena. The theory can 
be stated by means of the three fundamental and connected 
concepts of 'process', 'structure' and 'function'. It is derived 
from such earlier writers as Montesquieu, Comte, Spencer, 
Durkheim and thus belongs to a cultural tradition of two 
hundred years. This introduction contains a reformulation in 
which certain terms are used differently from the way they were 
used in the early papers here reprinted. For example, in the 
earliest papers written twenty or more years ago the word 'culture' 
is used in the accepted meaning of that time as a general term 
for the way of life, including the way of thought, of a particular 
locally defined social group. 



A MONGST primitive peoples in many parts of the world a 
/\ good deal of importance is attached to the relationship of 
^ jLmother's brother and sister's son. In some instances, the 
sister's~^n~has^ certain special rights over the property of his 
mother's brother. At one time it was usual to regard these customs 
as being connected with matriarchal institutions, and it was held 
that their presence in a patrilineal people could be regarded as 
evidence that that people had at some time in the past been matri- 
lineal. This view is still held by a few anthropologists and has 
been adopted by Mr. Junod in his book on the BaThonga people 
of Portuguese East Africa. Referring to the customs relating to 
the behaviour of the mother's brother and the sister's son to one 
another, he says: 'Now, having enquired with special care into this 
most curious feature of the Thonga system, I come to the con- 
clusion that the only possible explanation is that, in former and 
very remote times, our tribe has passed through the matriarchal 
stage.' (Junod, The Life of a South African Tribe, 191 3, Vol. I, 

p. 253-) 

It is with this theory that I wish to deal in this paper; but I 
do not propose to repeat or add to the objections that have been 
raised against it by various critics in recent years. Purely negative 
criticism does not advance a science. The only satisfactory way 
of getting rid of an unsatisfactory hypothesis is to find a better 
one. I propose, therefore, to put before you an alternative hypo- 
thesis, and if I am successful, not in proving my h3^o- 
thesis, but in showing that it does give a possible explanation ot 
the facts, I shall at least have refuted the view of Mr. Junod that 
the explanation he accepts is the 'only possible' one. 

For many African tribes we have almost no information about 
customs of this kind. Not that the customs do not exist, or 
are not important to the natives themselves, but because the 

^ A paper read before the South African Association for the Advancement 
of Science, 9 July 1924, and printed in South African Journal of Science, 
Vol. XXI, pp. 542-55. 


systematic and scientific study of the natives of this country has 
as yet hardly begun. I shall, therefore, have to refer chiefly to 
the customs of the BaThonga as recorded by Mr. Junod. These are 
to be found in the first volume of the work quoted above (pp. 225 
et seq., and pp. 253 et seq.). Some of the more important of them 
may be summarised as follows: 

1. The uterine nephew all through his career is the object of 
special care on the part of his uncle. 

2. When the nephew is sick the mother's brother sacrifices 
on his behalf. 

3. The nephew is permitted to take many liberties with his 
mother's brother; for example, he may go to his uncle's 
home and eat up the food that has been prepared for the 
latter's meal. 

4. The nephew claims some of the property of his mother's 
brother when the latter dies, and may sometimes claim 
one of the widows. 

5. When the mother's brother offers a sacrifice to liis ancestors 
the sister's son steals and consumes the portion of meat or 
beer offered to the gods. 

It must not be supposed that these customs are peculiar to 
the BaThonga. There is evidence that similar customs may be 
found amongst other African tribes, and we know of the existence 
of similar customs amongst other peoples in various parts of the 
world. In South Africa itself customs of this kind have been found 
by Mrs. Hoernle amongst the Nama Hottentots. The sister's son 
may behave with great freedom towards his mother's brother, and 
may take any particularly fine beast from his herd of cattle, or any 
particularly fine object that he may possess. On the contrary, the 
mother's brother may take from his nephew's herd any beast 
that is deformed or decrepit, and may take any old and worn-out 
object he may possess. 

What is particularly interesting to me is that in the part of 
Polynesia that I know best, that is, in the Friendly Islands 
(Tonga) and in Fiji, we find customs that show a very close 
resemblance to those of the BaThonga. There, also, the sister's 
son is permitted to take many liberties with his mother's brother, 
and to take any of his uncle's possessions that he may desire. And 
there also we find the custom that, w'hen the uncle makes a 


sacrifice, the sister's son takes away the portion offered to the gods, 
and may eat it. I shall, therefore, make occasional references to the 
Toogan customs in the course of this paper. 

' These three peoples, the BaThonga, the Nama, and the 
Tongans, have patrilineaLjor patriarchal institutions; that is, 
the children belong to the social group of the father, not to that 
of the mother; and property is inherited in the male line, passing 
normally from a father to his son^The view that I am opposing is 
that the customs relating to the mother's brother can only be 
explained by supposing that, at some past time, these peoples 
had matrilineal institutions, such as are found today amongst 
other primitive peoples, with whom the children belong to 
the social group of the mother, and property is inherited in the 
female line, passing from a man to his brother and to his 
sister's sons. 

llt^s a mistake to suppose that we can understand the insti- 
tutions of society by studying them in isolation without regard 
to other institutions with which they coexist and with which they 
may be correlated, and I wish to call attention to a correlation 
that seems to exist between customs relating to the mother's 
brother and customs relating to the father's sister. So far as 
present information goes, where we find the mother's brother 
important we also find that the father's sister is equally important, _ 
though in a different way. The custom of allowing the sister's' 
son to take liberties with his mother's brother seems to be 
generally accompanied with an obligation of particular respect 
and obedience to the father's sister.-Mr. Junod says little about 
the father's sister amongst the BaThonga. Speaking of a man's 
behaviour to this relative (his rarana) he says simply: 'He shows 
her great respect. However, she is not in any way a mother 
{mamanay (op. cit., p. 223). About the Nama Hottentots we 
have better information, and there the father's sister is the object 
of the very greatest respect on the part of her brother's child. 
In Tonga this custom is very clearly defined. A man's father's 
sister is the one relative above all others whom he must respect 
and obey. If she selects a wife for him he must marry her without 
even venturing to demur or to voice any objection; and so through- 
out his life, jjis fother^^i^teris^sao^^^ her word is his 
law; and one of the greatest offences of which he could be guilty 
would be to show himself lacking in respect to her. 


Now this correlation (which is not confined, of course, 
to the three instances I have mentioned, but seems, as I have said, 
to be general) must be taken into account in any explanation of 
the customs relating to the mother's brother, for the correlated 
customs are, if I am right, not independent institutions, but part 
of one system; and no explanation of one part of the system is 
satisfactory unless it fits in with an analysis of the system as 
a whole. 

, In most primitive societies the social relations of individuals 
ar^r-^ery largely regulated on the basis of kinship. This is brought 
about by the formation of fixed and more or less definite patterns 
of behaviour for each of the recognised kinds of relationship. 
There is a special pattern of behaviour, for example, for a son 
towards his father, and another for a younger brother towards 
his elder brother. The particular patterns vary from one society 
to another; but there are certain fundamental principles or 
tendencies which appear in all societies, or in all those of a 
certain type. It is these general tendencies that it is the special 
task of social anthropology to discover and explamT] 

Once we start tracing out relationship to any-xonsiderable 
distance the number of different kinds of relatives that it is 
logically possible to distinguish is very large. This difficulty is 
avoided in primitive society by a system of classification, by 
which relatives of what might logically be held to be of different 
kinds are classified into a limited number of kinds. The principle 
of classification that is most commonly adopted in primitive 
society may be stated as that of the equivalence of brothers. In 
other words if I stand in a particular relation to one man I regard 
myself as standing in the same general kind of relation to his 
brother; and similarly with a woman and her sister. In this way 
the father's brother comes to be regarded as a sort of father, and 
his sons are, therefore, relatives of the same kind as brothers. 
Similarly, the mother's sister is regarded as another mother, and 
her children are therefore brothers and sisters. The system is the 
one to be found amongst the Bantu tribes of South Africa, and 
amongst the Nama Hottentots, and also in the Friendly Islands. 
By means of this principle primitive societies are able to arrive 
at definite patterns of behaviour towards uncles and aunts and 
cousins of certain kinds. A man's behaviour towards his father's 
brother must be of the same general kind as his behaviour 



towards his own father and he must behave to his mother's sister 
according to the same pattern as towards his mother. The children 
of his father's brother or of the mother's sister must be treated 
in very much the same way as brothers and sisters. 

This principle, however, does not give us immediately any 
pattern for either the mother's brother or the father's sister. It 
would be possible, of course, to treat the former as being like a 
father and the latter as similar to a mother, and this course does 
seem to have been adopted in a few societies. A tendency in this 
direction is found in some parts of Africa and in some parts of 
Polynesia. But it is characteristic of societies in which the classi- 
ficatory system of kinship is either not fully developed or has 
been partly effaced. 

Where the classificatory system of kinship reaches a high 
degree of development or elaboration another tendency makes its 
appearance: the tendency to develop patterns for the mother's 
brother and the father's sister by regarding the former as a sort 
of male mother and the latter as a sort of female father. This 
tendency sometimes makes its appearance in language. Thus, 
in South Africa the common term for the mother's brother is 
malume or umalume, which is a compound formed from the stem 
for 'mother' — ma — and a suffix meaning 'male'. Amongst the 
BaThonga the father's sister is called rarana, a term which Mr. 
Junod explains as meaning 'female father'. In some South 
African languages there is no special term for the father's sister; 
thus in Xosa, she is denoted by a descriptive term udade ho hawo, 
literally 'father's sister'. In Zulu she may be referred to by a 
similar descriptive term or she may be spoken of simply as ubaba, 
'father', just like the father's brothers. In the Friendly Islands 
the mother's brother may be denoted by a special term tuasitia, 
or he may be called fu'e tangata, literally 'male mother'. This 
similarity between South Africa and Polynesia cannot, I think, 
be regarded as accidental; yet there is no possible connection 
between the Polynesian languages and the Bantu languages, and 
I find it very difficult to conceive that the two regions have adopted 
the custom of calling the mother's brother by a term meaning 
'male mother' either from one another or from one common 

Now let us see if we can deduce what ought to be the patterns 
of behaviour towards the mother's brother and the father's 


sister in a patrilineal society on the basis of the principle or 
tendency which I have suggested is present. I'o do this we must 
first know the patterns for the father and the mother respectively, 
and I think that it will, perhaps, be more reassuring if I go for 
the definition of these to Mr. Junod's work, as his observations 
will certainly not have been influenced by the hypothesis that I 
am trying to prove. 
\^ The relationship of father, he says, 'implies respect and even 
fear. The father, though he does not take much trouble with 
his children, is, however, their instructor, the one who scolds 
and punishes. So do also the father's brothers' (op. cit., p. 222). 
I Of a man's own mother he says: 'She is his true mamana, and this 
/ relation is very deep and tender, combining respect with love. 
Love, however, generally exceeds respect' (op. cit., p. 224). 
Of the mother's relation to her children we read that 'She is 
generally weak with them and is often accused by the father of 
spoiling them.' 

r There is some danger in condensed formulae, but I think 
we shall not be far wrong in saying that in a strongly patriarchal 
society, such as we find in South Africa, the father is the one w^ho 
must be respected and obeyed, and the mother is the one from 
whom may be expected tenderness and indulgence. I could show 
you, if it were necessary, that the same thing is true of the family 
life of the Friendly Islanders. 

If, now, we apply the principle that I have suggested is at 
work in these peoples it will follov/ that the father's sister is 
one who must be obeyed and treated with respect, while from the 
mother's brother indulgence and care may be looked for. But the 
matter is complicated by another factor. If we consider the relation 
of a nephew to his uncle and aunt, the question of sex comes in. 
In primitive societies there is a marked difference in the behaviour 
of a man towards other men and that towards women. Risking 
once more a formula, we may say that any considerable degree 
of familiarity is generally only permitted in such a society as the 
BaThonga between persaijs of the same sex. A man must treat 
his female relatives with greater respect than his male relatives. 
Consequently the nephew must treat his father's sister with even 
greater respect than he does his own father. (In just the same way, 
owing to the principle of respect for age or seniority, a man must 
treat his father's elder brother with more respect than his own 


father.) Inversely, a man may treat his mother's brother, who is of 
his own sex, with a degree of famiHarity that would not be possible 
with any w^oman, even his own mother. The influence of sex 
on the behaviour of kindred is best seen in the relations of brother 
and sister. In the Friendly Islands and amongst the Nama a man 
must pay great respect to his sister, particularly his eldest sister, 
and may never indulge in any familiarities with her. The same 
thing is true, I believe, of the South African Bantu. In many 
primitive societies the father's sister and the elder sisters are the 
objects of the same general kind of behaviour, and in some of 
these the two kinds of relatives are classified together and denoted 
by the same name. 

We have deduced from our assumed principle a certain pattern "^ 
of behaviour for the father's sister and for the mother's brother. 
Now these patterns are exactly what we find amongst the 
BaThonga, amongst the Hottentots, and in the Friendly Islands. 
The father's sister is above all relatives the one to be respected and 
obeyed. The mother's brother is the one relative above all from 
whom we may expect indulgence, with whom we may be familiar 
and take liberties. Here, then, is an alternative 'possible ex- 
planation' of the customs relating to the mother's brother, and 
it has this advantage over Mr. Junod's theory that it also explains 
the correlated customs relating to the father's sister. TKis brings 
us, however, not to the end but to the beginning of our enquiry. 
It is easy enough to invent hypotheses. The important and difficult 
work begins when we set out to verify them. It will be impossible 
for me, in the short time available, to make any attempt to verify 
the hypothesis I have put before you. All I can do is to point 
out certain lines of study which will, I believe, provide that 

The first and most obvious thing to do is to study in detail 
the behaviour of the sister's son and the mother's brother to 
one another in matriarchal societies. Unfortunately, there is 
practically no information on this subject relating to Africa, and 
very little for any other part of the world. Moreover, there are 
certain false ideas connected with this distinction of societies 
into matriarchal and patriarchal that it is necessary to remove 
before we attempt to go further. 

In all societies, primitive or advanced, kinship is necessarily 
bilateral. The individual is related to certain persons through his 


father and to others through his mother, and the kinship system 
of the society lays down what shall be the character of his dealings 
with his paternal relatives and his maternal relatives respectively. 
But society tends to divide into segments (local groups, lineages, 
clans, etc.), and when the hereditary principle is accepted, as it 
most frequently is, as the means of determining the membership 
of a segment, then it is necessary to choose between maternal 
or paternal descent. When a society is divided into groups with a 
rule that the children belong to the group of the father we have 
patrilineal descent, while if the children always belong to the 
group of the mother the descent is matrilineal. 

There is, unfortunately, a great deal of looseness in the 
use of the terms matriarchal and patriarchal, and for that reason 
many anthropologists refuse to use them. If we are to use them at 
all, we must first give exact definitions. A society may be called 
patriarchal when descent is patrilineal (i.e. the children belong 
to the group of the father), marriage is patrilocal (i.e. the wife 
removes to the local group of the husband), inheritance (of 
property) and succession (to rank) are in the male line, and jhe 
family is patripotestal (i.e. the authority over the members of the 
family is in the hands of the father or his relatives). On the 
other hand, a society can be called matriarchal when descent, 
inheritance and succession are in the female line, marriage is 
matrilocal (the husband removing to the home of his wife), and 
when the authority over the children is wielded by the mother's 

If this definition of these opposing terms is accepted, it 
is at once obvious that a great number of primitive societies are 
neither matriarchal nor patriarchal, though some may incline 
more to the one side, and others more to the other. Thus, if we 
examine the tribes of Eastern Australia, which are sometimes 
spoken of as matriarchal, we find that marriage is patrilocal, so that 
membership of the local group is inherited in the male line, the 
authority over the children is chiefly in the hands of the father and 
his brothers, property (what there is of it) is mostly inherited 
in the male line, while, as rank is not recognised, there is no 
question of succession. The only matrilineal institution is the . 
descent of the totemic group, which is through the mother, so 
that these tribes, so far from being matriarchal, incline rather 
to the patriarchal side. Kinship amongst them is thoroughly 


bilateral, but for most purposes kinship through the father is of 
more importance than kinship through the mother. There is some 
evidence, for example, that the obligation to avenge a death falls 
upon the relatives in the male line rather than upon those in the 
female line. 

We find an interesting instance of this bilateralism, if it may 
be so called, in South Africa, in the OvaHerero tribe. The facts 
are not quite certain, but it would seem that this tribe is sub- 
divided into two sets of segments crossing one another. For one 
set (the omaanda) descent is matrilineal, while for the other (otuzo) 
it is patrilineal. A child belongs to the eanda of its mother and 
inherits cattle from its mother's brothers, but belongs to the oruzo 
of its father and inherits his ancestral spirits. Authority over the 
children would seem to be in the hands of the father and his 
brothers and sisters. 

It is now clear, I hope, that the distinction between matriarchal 
and patriarchal societies is not an absolute but a relative one. 
Even in the most strongly patriarchal society some social im- 
portance is attached to kinship through the mother; and similarly 
in the most strongly matriarchal society the father and his kindred 
are always of some importance in the life of the individual. 

In Africa we have in the south-east a group of tribes that 
incline strongly to patriarchy, so much so, in fact, that we may 
perhaps justifiably speak of them as patriarchal. Descent of 
the social group, inheritance of property, succession to chieftain- 
ship, are all in the male line; marriage is patrilocal, and authority 
in the family is strongly patripotestal. In the north of Africa, in 
Kenya and the surrounding countries, there is another group of 
strongly patriarchal peoples, some of them Bantu-speaking, while 
others are Nilotic or Hamitic. Between these two patriarchal 
regions there is a band of peoples stretching apparently right 
across Africa from east to west, on the level of Nyasaland and 
Northern Rhodesia, in which the tendency is towards matriarchal 
institutions. Descent of the social group, inheritance of property, 
and succession to the kingship or chieftainship are in the female 
line. In some of the tribes marriage seems to be matrilocal, at any 
rate temporarily if not permanently, i.e. a man on marriage has to 
go and live with his wife's people. 

It is about these people and their customs that we urgently 
need information if we are to understand such matters as the 


subject of this paper. Of one tribe of this region we have a fairly 
full description in the vi^ork of Smith and Dale {The Ila-speaking 
People of Northern Rhodesia, 1920). Unfortunately, on the very 
points with which I am now dealing the information is scanty and 
certainly very incomplete. There are, however, two points I wish 
to bring out. The first concerns the behaviour of the mother's 
brother to his sister's son. We are told that 'the mother's brother 
is a personage of vast importance; having the power even of life 
and death over his nephews and nieces, which no other relations, 
not even the parents, have; he is to be held in honour even above 
the father. This is avunculi potestas, which among the Balla is 
greater than patria potestas. In speaking of the mother's brother, 
it is customary to use an honorific title given to people who are re- 
spected very highly' (op. cit.. Vol. I, p. 230). This kind of relation 
between the mother's brother and the sister's son is obviously 
what we might expect in a strongly matriarchal society. But how 
then, on Mr. Junod's theory, can we explain the change which 
must have taken place from this sort of relation to that which now 
exists among the BaThonga? 

This brings me to another point which it will not be possible 
to discuss in detail but which has an important bearing on the 
argument. We have been considering the relation of the sister's 
son to his mother's brother; but if we are to reach a really final 
explanation, we must study also the behaviour of a man to his 
other relatives on the mother's side, and to his mother's group 
as a whole. Now in the Friendly Islands the peculiar relation 
between a sister's son and a mother's brother exists also between 
a daughter's son and his mother's father. The daughter's son 
must be honoured by his grandfather. He is 'a chief to him. 
He may take his grandfather's property, and he may take away 
the ofi'ering that his grandfather makes to the gods at a kava 
ceremony. The mother's father and the mother's brother are the 
objects of very similar behaviour patterns, of which the out- 
standing feature is the indulgence on the one side and the liberty 
permitted on the other. Now there is evidence of the same thing 
amongst the BaThonga, but again we lack the full information 
that we need. Mr. Junod WTites that a grandfather 'is more lenient 
to his grandson by his daughter than his grandson by his son' 
(op cit. p. 227). In this connection the custom of calling the 
mother's brother kokwana (grandfather) is significant. 

THE mother's brother IN SOUTH AFRICA 2$ 

Now here is something that it seems impossible to explain 
on Mr. Ju nod's theory. In a strongly matriarchal society the 
mother's father does not belong to the same group as his grand- 
child and is not a person from whom property can be inherited 
or who can exercise authority. Any explanation of the liberties 
permitted towards the mother's brother cannot be satisfactory 
unless it also explains the similar liberties towards the mother's 
father which are found in Polynesia, and apparently to some extent 
in South Africa. This Mr, Junod's theoii^ clearly does not do, 
and cannot do. 

But on the hypothesis that I have put forw^ard the matter! 
is fairly simple. In primitive society there is a strongly marked 
tendency to merge the individual in the group to which he or shei 
belongs. The result of this in relation to kinship is a tendency 
to extend to all the members of a group a certain type of behaviour' 
which has its origin in a relationship to one particular member 
of the group. Thus the tendency in the BaThonga tribe would 
seem to be to extend to all the members of the mother's group 
(family or lineage) a certain pattern of behaviour which is derived 
from the special pattern that appears in the behaviour of a son i., 
towards his mother. Since it is from his mother that he expects I 
care and indulgence he looks for the same sort of treatment from 1 
the people of his mother's group, i.e. from all his maternal kin. y^ 
On the other hand it is to his paternal kin that he owes obedience 
and respect. The patterns that thus arise in relation to the.father_ 
and the mother are_generalised and extended to the kindred 
on the one side and on the other. If I had time I think I could 
show you quite conclusively that this is really the principle that 
governs the relations between an individual and his mother's 
kindred in the patriarchal tribes of South Africa. I must leave the 
demonstration, however, to another occasion. I can do no more 
now than illustrate my statement. 

The custom, often miscalled bride-purchase and generally 
known in South Africa asjo^o/^jis, as Mr. Junod has well shown, a 
payment made in compensation to a girl's family for her loss when 
she is taken away in marriage. Now, since in the patriarchal tribes- — 
of South Africa a woman belongs to her father's people, the com- 
pensation has to be paid to them. But you will find that in many 
j)i the tribes a certain portion of the 'marriage payment' is handed 
over to the mother's brother of the girl for whom it is paid. Thus, 



amongst the BaPedi, out of the lenyalo cattle one head (called 
hloho) is handed to the mother's brother of the girl. Amongst the 
BaSotho a portion of the cattle received for a girl on her marriage 
may sometimes be taken by her mother's brother, this being 
known as ditsoa. Now the natives state that the ditsoa cattle 
received by the mother's brother are really held by him on behalf 
of his sister's children. If one of his sister's sons or daughters is ill 
he may be required to offer a sacrifice to his ancestral spirits, and 
he then takes a beast from the ditsoa herd. Also, when the sister's 
son wishes to obtain a wife, he may go to his mother's brother to 
help him to find the necessary cattle and his uncle may give him 
some of the ditsoa cattle received at the marriage of his sister, or 
may even give him cattle from his own herd, trusting to being re- 
paid from the ditsoa cattle to be received in the future from the 
marriage of a niece. I believe that the Native Appeal Court has 
decided that the payment of ditsoa to the mother's brother is a 
voluntary matter and cannot be regarded as a legal obligation, and 
with that judgment I am in agreement. I quote this custom 
because it illustrates the sort of interest that the mother's brother 
is expected to take in his sister's son, in helping him and looking 
after his welfare. It brings us back to the question as to why the 
mother's brother may be asked to offer sacrifices when his nephew 
is sick. 

In south-east Africa ancestor worship is patrilineal, i.e. a 
man worships and takes part in sacrifices to the spirits of his 
deceased relatives in the male line. Mr. Junod's statements about 
the BaThonga are not entirely clear. In one place he says that 
each family has two sets of gods, those on the father's side and 
those on the mother's; they are equal in dignity and both can be 
invoked (op. cit., II, p. 349, and I, p. 256, note). But in another 
place it is stated that if an offering has to be made to the gods of 
the mother's family this must be through the maternal relatives, 
the malume (op. cit., II, p. 367). Other passages confirm this and 
show us that ancestral spirits can only be directly approached in 
any ritual by their descendants in the male line. 

The natives of the Transkei are very definite in their state- 
ments to me that a person's maternal gods, the patrilineal ancestors 
of his mother, will never inflict supernatural punishment upon 
him by making him sick. (I am not quite so sure about the Sotho 
tribes, but I think that they probably have similar views.) On 


the Other hand a married woman can receive protection from the 
ancestral spirits of her patrihneal Hneage, and so can her young 
children as long as they are attached to her. For children are 
only fully incorporated in their father's lineage when they reach 
adolescence. So in the Transkei a woman, when she marries, 
should be given a cow, the uhidiinga cow, by her father, from the 
herd of her lineage, which she can take to her new home. Since 
she may not drink the milk from her husband's herd during 
the early period of her married life she can be provided with 
milk from this beast that comes from her lineage. This cow 
constitutes a link between herself and her lineage, its cattle, and 
its gods, for cattle are the material link between the living members 
of the lineage and the ancestral spirits. So if she is sick she can 
make for herself a necklace of hairs from the tail of this cow and 
so put herself under the protection of her lineage gods. Moreover, 
if one of her infant children is sick, she can make a similar 
necklace which is thought to give protection to the child. When 
her son is grown up he should receive an uhulunga bull from his 
father's herd, and thereafter it is from the tail of this beast that 
he will make a protective amulet; similarly the daughter, when 
she marries, is detached from her mother, and may receive an 
iibulunga cow from her father. 

But though, according to the statements made to me, the 
maternal ancestors vrill not punish their descendant with sickness, 
they can be appealed to for help. When, therefore, a child is sick 
the parents may go to the mother's brother of the child, or to the 
mother's father if he is still living, and ask that a sacrifice shall 
be offered, and an appeal for help made to the child's maternal 
ancestors. This, at any rate, is stated as a practice in the Sotho' 
tribes, and one of the purposes of the ditsoa cattle that go from the 
marriage payment to the mother's brother of the bride is said to 
be to make provision for such sacrifices if they should be needed. 

This brings us to the final extension of the principle that' 
I have suggested as the basis of the customs relating to the mother's 
brother. The pattern of behaviour towards the mother, which is 
developed in the family by reason of the nature of the family 
group and its social life, is extended with suitable modifications to 
the mother's sister and to the mother's brother, then to the group 
of maternal kindred as a whole, and finally to the maternal gods, 
the ancestors of the mother's group. In the same way the pattern 


of behaviour towards the father is extended to the father's brothers 
and sisters, and to the whole of the father's group (or rather to all 
the older members of it, the principle of age making important 
modifications necessary), and finally to the paternal gods. 

The father and his relatives must be obeyed and respected 
(even worshipped, in the original sense of the word), and so 
therefore also must be the paternal ancestors. The father punishes 
his children, and so may the ancestors on the father's side. On 
the other hand, the mother is tender and indulgent to her child, 
and her relatives are expected to be the same, and so also the 

t.ernal spirits. 
A very important principle, which I have tried to demonstrate 
elsewhere {The Andaman Islanders, Chapter V), is that the social 
values current in a primitive society are maintained by being 
expressed in ceremonial or ritual customs. The set of values 
that we here meet with in the relations of an individual to his 
kindred on the two sides must, therefore, also have their proper 
ritual expression. The subject is too vast to deal with at all 
adequately here, but I wish to discuss one point. Amongst the 
BaThonga, and also in Western Polynesia (Fiji and Tonga), the 
sister's son (or in Tonga also the daughter's son) intervenes in the 
sacrificial ritual. Mr. Junod describes a ceremony of crushing 
down the hut of a dead man in which the batukulu (sister's chil- 
dren) play an important part. They kill and distribute the sacrificial 
victims and when the officiating priest makes his prayer to the 
spirit of the dead man it is the sister's sons who, after a time, 
interrupt oi* *cut' the prayer and bring it to an end. They then, 
among the BaThonga clans, seize the portions of the sacrifice 
that have been dedicated to the spirit of the dead man and run 
away with them, 'stealing' them (op. cit., I, p. 162).^^ 

I would suggest that the meaning of this iS'THat it gives a 
ritual expression to the special relation that exists between the 
sister's son and the mother's brother. When the uncle is alive 
the nephews have the right to go to his village and take his food. 
Now that he is dead they come and do this again, as part of the 
funeral ritual, and as it were for the last time, i.e. they come and 
steal portions of meat and beer that are put aside as the portion of 
the deceased man. 

The same sort of explanation will be found to hold, I think, 
of the part played in sacrificial and other ritual by the sister's 


son both amongst the Bantu of South Africa and also in Tonga 
and Fiji. As a man fears his father, so he fears and reverences his 
paternal ancestors, but he has no fear of his mother's brother, 
and so may act irreverently to his maternal ancestors; he is, 
indeed, required by custom so to act on certain occasions, thus 
giving ritual expression to the special social relations between a 
man and his maternal relatives in accordance with the general 
function of ritual, as I understand it. 

It will, perhaps, be of help if I give you a final brief statement 
of the hypothesis I am advancing, with the assumptions involved 
in it and some of its important implications. 

1. The characteristic of most of these societies that we 
call primitive is that the conduct of individuals to one 
another is very largely regulated on the basis of kinship, 
this being brought about by the formation of fixed patterns 
of behaviour for each recognised kind of kinship relation. 

2. This is sometimes associated with a segmentary organisation 
of society, i.e. a condition in which the whole society is 
divided into a number of segments (lineages, clans). 

3. While kinship is ahvays and necessarily bilateral, or 
cognatic, the segmentary organisation requires the adoption 
of the uniHneal principle, and a choice has to be made 
between patrilineal and matrilineal institutions. 

4. In patrihneal societies of a certain type, the special pattern 
of behaviour between a sister's son and the mother's 
brother is derived from the pattern of behaviour between 
the child and the mother, which is itself the product of the 
social life within the family in the narrow sense. 

5. This same kind of behaviour tends to be extended to all 
the maternal relatives, i.e. to the whole family or group to 
which the mother's brother belongs.^ 

^ This extension from the mother's brother to the other maternal relatives 
is shown in the BaThonga tribe in the kinship terminology. The term malunie, 
primarily applied to the mother's brother, is extended to the sons of those men, 
who are also malume. If my mother's brothers are dead it is their sons who will 
have to sacrifice on my behalf to my inaternal ancestors. In the northern part 
of the tribe the term malume has gone out of use, and the mother's father, 
the mother's brother, and the sons of the mother's brother are all called 
kokzvana (grandfather). However absurd it may seem to us to call a mother's 
brother's son, who may be actually younger than the speaker, by a word 
meaning 'grandfather', the argument of this paper will enable us to see some 
meaning in it. The person who must sacrifice on my behalf to my maternal 


6. In societies with patrilineal ancestor worship (such as 
the BaThonga and the Friendly Islanders) the same type of 
behaviour may also be extended to the gods of the mother's 

7, The special kind of behaviour to the maternal relatives 
(living and dead) or to the maternal group and its gods and 
sacra, is expressed in definite ritual customs, the function 
of ritual here, as elsewhere, being to fix and make permanent 
certain types of behaviour, with the obligations and senti- 
ments involved therein. 

In conclusion, may I point out that I have selected the subject 
of my contribution to this meeting because it is one not only of 
theoretical but also of practical interest. For instance, there is the 
question as to whether the Native Appeal Court was really right 
in its judgment that the payment of the ditsoa cattle to the mother's 
brother of a bride is not a legal but only a moral obligation. So 
far as I have been able to form an opinion, I should say that 
the judgment was right. 

■ The whole subject of the payments at marriage [lohola) is 
/one of considerable practical importance at the present time to 
missionaries and magistrates, and to the natives themselves. Now 
the study of the exact position in which a person stands to his 
maternal relatives is one without which it is impossible to arrive 
at a completely accurate understanding of the customs of lobola. 
One of the chief functions of lohola is to fix the social position of 
the children of a marriage. If the proper payment is made by a 
family, then the children of the woman w^ho comes to them in 
exchange for the cattle belong to that family, and its gods are their 
gods. The natives consider that the strongest of all social bonds is 
that between a child and its mother, and therefore by the ex- 
tension that inevitably takes place there is a very strong bond 
between the child and its mother's family. The function of the 
lohola payment is not to destroy but to modify this bond, and to 
place the children definitely in the father's family and group for 
all matters concerning not only the social but also the religious 

ancestors is first my mother's father, then, if he is dead, my mother's brother, 
and after the decease of the latter, his son, who may be younger than I am. 
There is a similarity of function for these three relationships, a single general 
pattern of behaviour for me towards them all and this is again similar in general 
to that for grandfathers. The nomenclature is, therefore, appropriate. 


life of the tribe. If no lobola is paid the child inevitably belongs 
to the mother's family, though its position is then irregular. 
But the \A Oman for whom the lobola is paid does not become a 
member of the husband's family; their gods are not her gods; 
and that is the final test. I have said enough, I hope, to show that 
the proper understanding of customs relating to the mother's- 
brother is a necessary preliminary to any final theory of lobola. 



IF we are to understand aright the laws and customs of non- 
European peoples we must be careful not to interpret them in 
terms of our own legal conceptions, which, simple and obvious 
as some of them may seem to us, are the product of a long and 
complex historical development and are special to our own culture. 
If, for instance, we attempt to apply to the customs of the simpler 
peoples our own precise distinctions between the law relating to 
persons and the law relating to things we shall produce nothing 
but confusion in the result. 

With us one of the most important aspects of succession is 
the transmission of property by inheritance. Yet in some of the 
simplest societies this is a matter of almost no significance at all. 
In an Australian tribe, for example, a man possesses a few weapons, 
tools, utensils and personal ornaments, things of little value or 
permanence. On his death some of them may be destroyed, others 
may be distributed among his relatives and friends. But their 
disposal is of so little importance, unless in relation to ritual, 
that it is often difficult to find any rules of customary procedure. 
But even in such simple societies, where inheritance of private 
property may be said not to exist or to be of minimal importance, 
there are problems of succession in the widest sense of the term. 
The term 'succession' will here be taken as referring to the 
transmission of rights in general. A right exists in, and is de- 
finable in terms of, recognised social usage. A right may be 
that of an individual or a collection of individuals. It may be 
defined as a measure of control that a person, or a collection of 
persons, has over the acts of some person or persons, said to be 
thereby made liable to the performance of a duty. Rights may be 
classified as of three main kinds: 

(a) Rights over a person imposing some duty or duties upon 

that person. This is the jus in personam of Roman law. 

A father may exercise such rights over his son, or a nation 

over its citizens. 

' Reprinted from The lozva Lazv Reviezv, \'ol. XX, No. 2, January 1935. 



(b) Rights over a person 'as against the world', i.e. imposing 
duties on all other persons in respect of that particular 
person. This is the jus in rem of Roman law in relation to 

{c) Rights over a thing, i.e. some object other than a person, 
as against the world, imposing duties on other persons in 
relation to that thing. 

The rights classified under {b) and (c) are fundamentally of 
the same kind, distinguished only as they relate to persons or to 
things, and are of a different kind from those classified under {a). 

We may consider a few examples from such a simple society 
as an Australian tribe. A man has certain rights over his wife. 
Some of them are rights in personam whereby he may require from 
her the performance of certain duties. Others are rights in rem. 
If anyone should kill the wife he commits an injury against the 
husband. If anyone should have sexual intercourse with the wife 
without the consent of the husband he commits an injury against 
the latter. In some tribes a man may lend his wife to another; this 
is an exercise by the husband both of his rights in personam and of 
his rights in rem. 

In a great number of Australian tribes the custom of the 
levirate holds sway. By this, when a man dies all his rights over 
his wife (and over his immature children) are transferred to his 
younger brother, or failing such, to an agnatic cousin. This is 
a simple instance of fraternal succession. What is transferred is 
certain rights in perso?iam and in rem over certain persons (the wife 
and children) and with these rights there go, of course, certain 
obligations or duties. 

Let us next consider, in such a tribe as the Kariera of Western 
Australia, the nature of the group that I shall call a 'horde'. 
iTfiis is a body of persons who jointly possess, occupy and'^p^loiTa-^ 
certain defined area of cotrrltryl The rights of the horde over its 
territory can be briefly indicStM by saying that no person who is 
not a member of the horde has the right to any animal, vegetable 
or mineral product from the territory except by invitation or 
consent of members of the horde. Acts of trespass against this 
exclusive right of a horde to its territory seem to have been very 
rare in the social life of the aborigines but it appears to have been 
generally held that anyone conmiitting such a trespass could 



justifiably be killed.^ This exclusive use of its territory by a horde 
is modified by obligations of hospitality whereby, when there is an 
abundance of some kind of food at a certain time, members of 
friendly neighbouring hordes are invited to come in and share it. 
The son of any woman born in the horde and married elsewhere 
is always^-entitled to visit his mother's horde and hunt in its 
territory. / 

It is^convenient to speak of such a group as the Kariera horde 
as a 'corporation' having an 'estate'. This is an extension of the 
terms 'corporation' and 'estate' as they are commonly used in law, 
but I think this extension is justifiable, and hope that at any rate 
it will be admitted for the purposes of the present exposition. By 
an estate is here meant a collection of rights (whether over persons 
or things) with the implied duties, the unity of which is constituted 
either by the fact that they are the rights of a single person and 
can be transmitted, as a whole, or in division, to some other 
person or persons, or that they are the rights of a defined group 
(the corporation) which maintains a continuity of possession. A 
personal estate thus corresponds to that universitas juris which is 
what, in Roman law, was transmitted by inheritance. ^ . . . 

The corporate estate of a Kariera horde includes in the first 
place its rights over its territory. The continuity of the horde 
is maintained by the continuity of possession of the territory, 
which remains constant, not subject to division or increase, for 
the Australian aborigines have no conception of the possibility 
of territorial conquest by armed force. The relation of a horde 
to its territory does not correspond exactly to what we regard as 
'ownership' in modern law. It has some of the qualities of corpor- 
ate ownership, but also partakes of the nature of the relation of 
a modern state to its territory, which we may speak of as the 
exercise of 'dominion'. Rights of ownership over land and rights 
of dominion have seemingly both had their origin by development 
and differentiation from such a simple relation as that exemplified 
in the Australian horde. 

^ We have records from a part of South Australia of occasional deliberate 
acts of trespass with armed force, a body of men invading a territory in which 
red ochre was found for the purpose of obtaining a supply. This was actually 
an act of war and as the invaders took care to come in force the horde whose 
rights were thus invaded had no effective remedy. 

^ Hereditas est successio in universum jus quod defunctus habuit: an inheritance 
is a succession to the entire legal position of a deceased man. 


The estate of a horde includes not only its rights over a 
territory, but also its rights, in personam and in rem, over its 
members. The adult male members of the horde owe certain 
duties to it so that it has rights in personam over them. It also has 
rights in rem, for if one of them is killed, by violence or by sorcery, 
the horde as a whole conceives itself to have been injured and 
takes steps to obtain satisfaction. Women and children are not 
members of the horde in the same sense as adult males. If a 
man's wife is 'stolen' it is he as an individual who seeks satisfaction 
though he will have the backing of the other members of the horde. 
But indirectly she also belongs to the horde so that when her 
husband dies she should by custom pass into the possession of some 
other member of the horde and not to some person outside. 

Since the Kariera horde is exogamous every female child 
passes by marriage out of the possession of her parents and out of 
the possession of the horde into the possession of her husband in 
another horde. By Australian custom this transfer of possession, 
i.e. of rights, in personam and in rem, over a person, should 
normally involve compensation or indemnification, which, in many 
tribes, is provided by the man who receives a wife giving his 
sister in exchange to be the wife of his brother-in-law. Male 
children may be said to pass out of the possession of the parents 
and into direct possession of the horde at initiation. This is, 
in some tribes, symbolically expressed in the initiation ritual. 

The Kariera horde affords an example of perpetual corporate 
succession. It will be obvious, I think, that it contains the germs 
of the state and of sovereignty as we know them in more complex 
developments. Thus, as the terms have been used above, the 
United States of America is a 'corporation' having as its con- 
stituent 'estate' possession of, or dominion over, a certain territory 
(subject, unlike that of an Australian horde, to increase by con- 
quest or purchase) and certain specific rights, in personam and 
in rem, over the persons of its citizens. 

The continuity of a corporation such as the Australian horde 
is dependent on the continuity of its estate. In the first place 
there is continuity of possession of the territory. Secondly, there 
is a continuity which transcends the space of a human life by the 
fact that as the group loses some members by death it acquires 
new members by the birth of children and the initiation of boys 
into the status of men. 


If now we turn from considering the horde as a whole to 
consider the individual male members we find here also a process 
of customary transmission of rights. Children 'belong', we may 
say, primarily to the father, i.e. it is he who exercises over them 
rights in personam and in rem. As the father in turn belongs to 
the horde this horde has some rights over his children. When a 
girl reaches puberty the rights over her are transferred (perhaps 
not in entirety but in great part) from her father and his horde to 
her husband. When the boy reaches puberty he is transferred from 
his position of dependence on his father to that of an adult member 
of the horde. Now a member of the horde has certain rights over 
other members and over the territory of the horde. These rights are 
part of his personal estate or status. Thus there is a process of 
'patrilineal succession' whereby the sons of male members of the 
horde become in their turn members, thus acquiring rights and 
having a share in the estate. 

We are thus brought, after necessary and it is hoped not too 
tedious preliminary considerations, to the problem with which this 
paper is to deal, that of the nature and function of the unilineal 
transmission of rights. In the patrilineal succession of the 
Australian horde the most considerable part of the body of rights 
of a male person, his status, his personal estate as a sharer or 
co-parcener in the estate of a horde, are derived by liim through 
his father to the exclusion of his mother and are transmitted in 
turn to his sons to the exclusion of his daughters. It is important, 
however, to recognise that in this instance, and, so far as we know, 
in all instances of patrilineal succession, some rights are also 
transmitted through the mother. Thus in the Kariera tribe a man 
has certain quite important rights over his mother's horde, over 
its individual members, and over its territory. 

In matrilineal succession the greater part of the body of 
rights of an individual, over things, over persons, or as a member 
of a corporation, are derived by him through his mother and can- 
not be transmitted to his children but devolve upon his sister's 

As an example of a very thorough system of matrilineal succes- 
sion we may consider the taravad of the Nayar caste of Malabar. 
A taravad is an incorporated matrilineal lineage. It includes all 
living descendants in the female line of an original ancestress. 
It has therefore both male and female members, all of whom are 


children of female members of the group. It is constituted as a 
corporation (a joint-family in the terminology of Indian lawyers) 
by the possession of an estate which includes in the first instance 
possession of a house or houses and land, and in the second place 
rights over the persons of its members. The control of the estate 
is in the hands of a 'manager' who is normally the oldest male 
member of the group. In order that the group may retain complete 
and exclusive possession of the children born to its female members 
the Nayars have established a system which denies all legal rights 
to a male parent. A Nayar girl is 'married' while still very young 
to a suitable bridegroom by the Hindu religious ceremony of 
the tying of a jewel. (It is probable that in former times the 
'bridegroom' ceremoniously deflowered the virgin 'bride'.) On 
the third day the newly-w^edded pair are divorced by the Hindu 
ceremony of dividing a piece of cloth. Thereafter the divorced 
bridegroom has no rights over the person, the estate or the children 
of his bride. At a later period the girl takes a lover. In former times, 
amongst some of the Nayar, if not generally, a woman was per- / 
mitted by custom to have two or more lovers at the same time. / 
As the lover is not married to the woman he also has no legal -j 
claims over her person or estate or over any children that may 
be born of the union. 

The Nayar system is the most thoroughgoing example of 
perpetual matrilineal succession. The lineage group maintains 
its continuity and its unity by not admitting any outside person 
to any share in its estate. It retains possession of its own women 
and claims exclusive rights over the children born to them. 

The status^ of an individual at a given moment of time may 
be defined as the totality of all his rights and duties as recognised 
in the social usages (laws and customs) of the society to which he 
belongs. The rights constituting a status, and similarly the duties, 
are of many different kinds, some relating to 'the world at large', 
to the society as a w-hole, others relating to some definite social 
group of which the individual is a member (e.g. a man's rights 
over and duties towards his ow^n clan), or to some group of which 
he is not a member but to which he stands in a special rela- 
tion (e.g. a man's relation to his mother's clan in a patrilineal 
clan system, or to his father's clan in a matrilineal system), and 

^ It is well always to remember that status, estate, state, and the French etat 
are all different forms of one and the same word, the late Latin estatus. 


yet others concern his special relations as an individual with 
other individuals. 

Everywhere in human society the status of an individual is 
very largely determined by birth as the child of a particular father 
and particular mother. Behind the question of succession, there- 
fore, lies the question of what elements of status, i.e. what rights 
and duties, are transmitted to the child by the father on the one 
hand and by the mother on the other. Every society has to establish 
its system of rules in this matter and there is an immense diversity 
of systems to be found in surviving and historic communities. The 
almost universal rule is that an individual derives some elements 
of his status from or through his father, and others from or through 
his mother. 

It has to be remembered that in all societies there is a general 
difference between the status of a man and that of a woman, 
and in some societies these differences are very marked and very 
important. Thus when a son 'succeeds' his father he may attain 
a status very similar to that of his father, but a daughter cannot do 
so to the same extent. The reverse holds true in the instance of 
a mother and her daughter on the one hand and her son on the 
other. Thus in African kingdoms w^here succession is matrilineal a 
king is succeeded by his younger brother and then by his sister's 
son. The heir therefore acquires, through his mother, important 
elements of the status of his mother's brother. The king's sister, 
who holds a very important position, is, of course, succeeded 
by her daughter. 

One solution of the problem of the determination of status 
would be to let the sons derive from the father and daughters 
from the mother. This principle is only known to be adopted in a 
few tribes, about which we know very little, in East Africa and in 
New Britain. As a working arrangement it has weighty objections 
which cannot be gone into here. 

It is possible to have a system in which a child, by birth, 
acquires the same rights, of the same kind and to an equal degree, 
over the persons to whom he is related through his father and those 
to whom he is related through his mother. An instance of this is 
that where a person has an equal expectation of testamentary or 
intestate succession to the estate of the brothers and sisters of 
his father and those of his mother. A further instance is provided 
by the customs relating to wergild amongst the Teutonic peoples. 


By birth a man acquired rights over a number of persons who 
constituted his sib.^ This included all his relatives through 
his father and through his mother, counting either through males 
or through females, within a certain range. This range varied 
in different Teutonic communities and perhaps in the same com- 
munity at different times. Amongst some of the Anglo-Saxons it 
extended as far as fifth cousins. If a man were killed all the members 
of his sib could claim a share in the indemnity (wergild) paid by the 
killer, proportionate to the degree of the relationship. Inversely, 
if a man killed another all members of his sib were under obligation 
to contribute to the blood-money he had to pay, each contributing 
in the same proportion as he would receive if the man himself 
had been killed. The members of a man's sib had specific rights 
in rem in relation to him and specific duties in personam towards 

The solution adopted by the great majority of human societies 
of the problem relating to the determination of status has been 
one by which a child derives certain rights and duties through 
the father and others of a different kind through the mother. Where 
the rights and duties derived through the father preponderate 
in social importance over those derived through the mother we 
have what it is usual to call a patrilineal system. Inversely a 
matrilineal system is one in which the rights and duties derived 
through the mother preponderate over those derived through the 

There are, however, some societies in which there is a fairly 
even balance between the elements of status derived through the 
father and those through the mother. An example is provided 
by the OvaHerero of south-west Africa. Through his mother a 
child derives membership in an eanda, a matrilineal clan; through 
his father he becomes a member of an oruzo, a patrilineal clan. 
There is thus a double system of clans crossing one another. 
As both kinds of clans are exogamous a man cannot belong to the 
eanda of his father or to the onizo of his mother. Through his 
mother and as a member of her eanda he has certain rights over, 
and duties towards, that group and particularly his mother's 

^ Professor Lowie and some American writers use the term 'sib' as equivalent 
to what is here, in accordance with European usage, called a 'clan'. It seems 
desirable to retain the word 'sib' for the bilateral group of kindred to which it 
was originally applied. 


l^rothers and his sister's children. Secular property is inherited 
only within the eanda so that a man inherits such property from 
his mother's brother and transmits it to his sister's son. On the 
other hand through his father and as a member of his oruzo he 
has rights and duties of other kinds in relation to that group. 
Certain sacred cattle may only be inherited within the onizo and 
are therefore transmitted from father to son. 

There are to be found in Africa and in Oceania other in- 
stances of systems in which patrilineal and matrilineal succession 
are combined and more or less balanced against one another. In 
a considerable part of Africa this is rationalised by a conception 
that every human being is compounded of two principles, one, 
called the 'blood' in Ashanti, derived from the mother, the other, 
the 'spirit', derived from the father. 

Probably the most important factor in determining the nature 
of succession in the simpler societies is the need of defining 
rights i7i rem over persons. When a child is born there is the 
question, 'To whom does the child belong?' It may, of course, be 
regarded as belonging jointly to the two parents. Both have an 
interest in it, both have rights in personam and ifi rem in relation 
to it. But there are other persons who have rights, in personam and 
i?i rem over the father (his parents, and brothers and sisters) and 
others who similarly have rights over the mother. In any society 
in which kinship is of fundamental importance in the total social 
structure, as it is in the majority of non-European societies, it is 
essential for social stability and continuity that the rights of 
different individuals over a given individual should be defined in 
such a way as to avoid as far as possible conflicts of rights. We 
have seen how the ancient Teutonic system gives similar, and in 
some instances equal, rights in rem to the father's kin and to the 
mother's kin of a given individual so that if he is killed all members 
of liis sib (i.e. liis kindred on both sides) are entitled to compen- 
sation. Let us now consider examples of the solution of this prob- 
lem in matrilineal and patrilineal systems. 

For a matrilineal system we may return to the Nayar as 
affording an extreme, and therefore crucial instance. In that 
system the taravad, or joint-family, maintains intact and absolute 
its rights in rem over all its members. Marriage normally gives the 
husband certain rights in rem over his wife and over the children. 
The Nayars may be said either to have eliminated marriage or to 


have eliminated this aspect of marriage. It is true that the union 
of a Nayar woman and her samhatidham lover is often a life-long 
union of great affection and that the lover has a great attachment 
to the children. But legally he has no rights over his 'wife', if 
we call her such, or over the children. In turn the group has no 
rights in rem over him for these remain with his own taravad. 
The taravad as a corporation retains undivided and undisputed 
possession of its own estate. 

As an instance of a definitely patrilineal solution of the problem 
of the distribution of rights in rem we may take the Zulu-Kaffir 
tribes of South Africa. In these tribes marriage requires the 
payment of an indemnity in the form of a number of cattle, 
called the Ikazi, the act of transfer of these being known as 
iiku-lobola. An unmarried girl belongs to her father, or to her 
guardian (father's brother or brother) if her father be dead, and 
to her agnatic kindred. They have over her certain rights in 
perso7iam and m rem. An offence committed against her, as rape, 
seduction, maiming or homicide, is an injury to her kin and they 
have the right to be indemnified. A father may bring before the 
chief an action for compensation for an offence committed against 
his daughter. By the act of marriage the father and the agnatic 
kindred surrender a great part of these rights over the daughter 
to her husband and to his agnatic kindred. The payment of cattle 
is an indemnification for this surrender of rights. For these 
people the great value of a woman is as the mother of children. 
(For this reason there is no more unhappy, unwanted person 
among them than a barren woman.) The act of lobola is therefore 
primarily a procedure whereby those paying the cattle acquire 
undivided and indisputable rights over all children born to the 
woman. This is readily demonstrated by an analysis that would 
be out of place here. The natives state the principle in two ways: 
'Cattle beget children'; 'The children are where the cattle are not.' 
In case of divorce either the wife and her children return to her 
father and any cattle paid are returned, or if (as is usual) the husband 
retains the children he must abandon claim to the cattle he has 
paid or to some portion thereof. On the death of a wife who has 
borne children (a barren wife may be repudiated and the re- 
payment of the cattle or the substitution of a sister may be claimed) 
if all the cattle have been paid, the children remain with the father 
and the mother's kin have no rights in rem over them. The system 


here outlined is a simple legal procedure for giving the father and 
his agnatic kindred indisputable and undivided rights in rem 
over his children. 

VThus the system of patrilineal or matrilineal succession 
cemfeTTargely around the system of marriage. In an extreme 
matrilineal society a man has no rights in rem over his children, 
though he does usually have certain rights in personam. The 
rights remain with the mother and her relatives. The result is to 
emphasise and maintain a close bond between brother and sister 

i at the expense of the bond between husband and wife. Con- 

( sequently the rights of the husband over his wife are limited. 
In an extreme patrilineal society we have exactly the opposite. 
Rights in rem over the children are exclusively exercised by the 
father and his relatives. The bond between husband and wife is 
strengthened at the expense of the bond between brother and 
sister. The rights of the husbajid over his wife are considerable; 

j she is in manu, under his potestasl 

I Extreme patrilineal systems^ are comparatively rare, and 

extreme matrilineal systems perhaps even rarer. Generally there 
is some modification by which, while the kindred on the one side 
have a preponderant right, some rights are recognised on the other 
side also. Thus in the Cherokee tribe of North American Indians, 
while a man belonged to his mother's clan, so that if he were 
killed they and they alone would demand satisfaction, yet he stood 
in a very special relation to his father and to his father's clan. 

Little has been said so far about the inheritance of property. 
This is because in the simpler societies the transmission of property 
is generally dependent upon the transmission of status. Thus 
amongst the Nayar the important property (land, houses, etc.) is 
the undivided or joint possession of a corporation constituted by a 
matrilineal lineage. Amongst the Zulu- Kaffir tribes the sons suc- 
ceed to a share of the father's estate to the exclusion of daughters 
y and their descendants. In general, though there are a few excep- 
tions, it may be said that the transmission of property follows the 
same line as does the transmission of status. 

With regard to the institutions of patrilineal and matrilineal 
succession the question is frequently asked as to what is their 
origin. The term 'origin' is ambiguous. In one sense we may talk 
of the 'historical origin'. The historical origin of the Nayar system, 
or that of the Zulu- Kaffirs, or of any other system, is a series of 


unique events extending often over a long period of gradual 
growth. The determination of the origin in this sense of any social 
system is the task of a historian. For the simpler peoples these 
histories are unknown and are the subject only of pure speculation, 
to my mind largely unprofitable. But the term 'origin' may be 
used in another sense, and very frequently it is used ambiguously 
with. a. confusion of the two meanings. "^- 

Any social system, to survive, must conform to certain con- 
ditions. If we can define adequately one of these universal 
conditions, i.e. one to which all human societies must conform, we 
have a sociological law. Thereupon if it can be shown that a 
particular institution in a particular society is the means by which 
that society conforms to the law, i.e. to the necessary condition, 
we may speak of this as the ' sociological origin' of the institution. 
Thus an institution may be said to have its general raison d'etre 
(sociological origin) and its particular raison d'etre (historical 
origin). The first is for the sociologist or social anthropologist 
to discover by the comparative method. The second is for the 
historian to discover by examination of records or for the eth- 
nologist, in the absence of records, to speculate about. j^y^ 

One such law, or necessary condition of continued existence, 
is that of a certain degree of functional consistency amongst the 
constituent parts of the social system. Functional consistency is not 
the same thing as logical consistency; the latter is one special 
form of the former. Functional inconsistency exists whenever two 
aspects of the social system produce a conflict which can only be 
resolved by some change in the system itself. It is always a question 
of the functioning, i.e. the working of the system as a whole. 
Consistency is a relative matter. No social system ever attains to 
a perfect consistency, and it is for this reason that every system 
is constantly undergoing change. Any insufficiency in this respect 
in a social system tends to induce change, sometimes, though by 
no means always, through the conscious recognition of the 
insufficiency by members of the society' and the conscious seeking 
of a remedy. To this law of the necessity of a certain degree of func- 
tional consistency we may add a second, which is a special instance 
of the first. Any human social life requires the establishment of 
a social structure consisting of a network of relations between 
individuals and groups of individuals. These relations all involve 
certain rights and duties which need to be defined in such a way 


that conflicts of rights can be resolved without destroying the 
structure. It is this need that is met by the establishment of 
systems of justice and legal institutions. 

Every system of rights necessarily involves the existence of 
common, joint or divided rights over the same person or thing. 
The father and the mother of a child both have rights in personam 
over their child. In an orderly family it is necessary that there should 
be no unresolved or unresolvable conflict between these rights. 
The same thing is true throughout a society as a whole in all the 
various relations into which persons are brought. When two 
persons A and B have rights over something Z or rights in rem 
over some person Z, there are three ways of adjusting these rights 
so as to avoid unresolvable conflicts. One is the mode of rights in 
common ; A and B have similar and equal rights over Z and these 
are such that the rights of A will not conflict with those of B. 
An instance is to be found in the native tribes of South Africa in 
which, as the native saying is, 'grass and water are common'. 
Any member of a tribe has the right to graze his cattle or water 
them, or take water for his own use, in any part of the territory 
over which the tribe (represented by its chief) exercises dominion. 
A second is the mode of joint rights in which A and B (or any 
number of persons) exercise jointly certain rights over Z. The 
establishment of such joint rights immediately establishes what is 
here called a corporation. An infringement of the rights normally 
calls for a joint action on the part of the corporation, which may, 
of course, be carried out by its official representatives. A South 
African tribe has joint possession of its territory, the possession 
(the estate) being vested in the chief. An infringement of these 
rights may be adjusted by the chief or may lead to the action of 
war, in which, under the chief, the whole tribe seeks to maintain 
its rights. The third mode is that of rights in division. Here A has 
certain definite rights over Z and B has certain other definite 
rights; the respective rights may be defined either by custom or by 
a specific contract, or agreement. An example is the relation of 
owner and tenant of a leased land or building. 

So far as rights over persons go the exercise of rights in 
common is necessarily very limited. In an unfamiliar region one 
may ask direction from any person one meets and expect to receive 
whatever information that person can give. In English law the 
king's officers can demand from any passer-by 'in the king's name' 


assistance in the arrest of a malefactor.^ Rights over persons 
in personam are usually exercised either jointly or in division. 

Rights over persons in rem can obviously never be exercised 
in common. We have seen, from the example of the Teutonic 
customs relating to wergild, that they can be held in division. But 
such a thing is rare and for the reason that it requires a compli- 
cated definition of the respective shares of various kindred in 
their interest in their kinsmen. One has only to glance at some 
of the early laws of Norway and Sweden relating to the division of 
the wergild between agnatic and cognatic kin of different degrees 
to realise with what difficulties such a system is confronted in its 

It results from this that rights in rem over a person must 
as a general rule either be exclusively personal, i.e. confined to a 
single individual (a condition to which the rights of an owner 
over a slave may approximate in some instances) or must be 
joint. The rights of a Roman father over his children were nearly 
exclusive but even these, at certain periods of history certainly, 
were subject to the rights, exercised jointly, of the gens or of the 
state; even the potestas of a pater familias was not absolute. Thus 
we may say that any society that recognises rights in rem over 
persons (and all known societies do so to some extent) will 
normally, and with only the rarest exceptions, make some provision 
for the joint exercise of such rights. This implies the existence 
of corporations of some kind, since a corporation is here defined as 
a collection of persons who jointly exercise some right or rights. 

A corporation can only form itself on the basis of a common 
interest. In the simplest societies the easiest, perhaps almost the 
only, ways in which common interests can be created are on 
the basis of locality, i.e. residence in the same local community 
or neighbourhood, or kinship. Corporations therefore tend to be 
established either on the one basis or the other or on both combined 
(the Kariera horde is an example of the latter) or else a double 
system of local groups and kinship groups is formed. 

We must here appeal to another sociological law, the necessity v' 
not merely for stability, definiteness and consistency in the social 
structure, but also for continuity. To provide continuity of social 

^ The latter instance, however, might be interpreted as the exercise of a 
joint right, since the king is the representative of the nation which, as a cor- 
poration, has joint rights over the persons of its citizens. 


Structure is essentially a function of corporations. Thus a modern 
nation has its continuity as a corporation exercising joint rights 
over its territory and over the persons of its citizens. 

We can imagine as a possibility an incorporated local com- 
munity which was completely endogamous and which would 
therefore not have to face the issue of choosing between matrilineal 
and patrilineal succession, since any child born in the community 
would have both its parents there. But the moment there are 
intermarriages between two corporate local groups the question of 
lineal succession does arise. In such a situation it is possible that 
no customary rule may be established, each instance being ad- 
justed by agreement of the persons most nearly concerned. It seems 
that this was the case of the hordes or local groups of the Andaman 
Islanders. The result is to produce a loose and indefinite structure. 
If any definite rule does arise it must usually take the form of one 
either of matrilineal or of patrilineal succession. 

If any society establishes a system of corporations on the basis 
of kinship — clans, joint-families, incorporated lineages — it must 
necessarily adopt a system of unilineal reckoning of succession. 
It would, of course, be theoretically possible to establish some sort 
of rule whereby, when the parents belong to diff"erent groups, in 
certain definite circumstances the children belong to the father's 
group and in others to the mother's. This would produce com- 
plicated conditions, and in general any complicated definition of 
rights is likely to be functionally inefficient as compared with a 
simpler one. 

Thus the existence of unilineal (patrilineal or matrilineal) 
succession in the great majority of human societies can be traced 
to its sociological 'cause' or 'origin' in certain fundamental 
social necessities. Chief amongst them, I have suggested, is the 
need of defining, with sufficient precision to avoid unresolvable 
conflicts, the rights in rem over persons. The need of precise 
definition of rights in personam and of rights over things would 
seem to be secondary but still important factors. 

There are many facts which might be adduced to support this 
hypothesis. I will mention only one kind. In societies organised 
on the basis of clans one of the most important activities of the 
clan is to exact vengeance or indemnification when a clansman is 
killed. The list of known instances of this would fill many pages. 
The clan as a corporation has rights in rem over all its clansmen. 


If one is killed the clan is injured and it has the right, and its 
members are under an obligation, to proceed to some action 
towards receiving satisfaction, either through vengeance or by 
receiving an indemnity. 

Thus the cause of the decay of the clan (the genos or gens) 
in ancient Greece and Rome was the transfer of its rights in rem 
(and necessarily therefore of some of its rights in personam ) to 
the city or state, the nature of these rights being inevitably 
considerably modified in the process of transference. But the 
decay of the gens in Rome still left the patriarchal family as a 
corporation (as Maine long ago pointed out) the basis of which, 
however, was not merely the exercise of rights in rem by the pater 
familias over his children, but also the exercise of joint rights 
over property and the maintenance of a religious cult of ancestor- 

/The sociological laws, i.e. the necessary conditions of exist- / 
ence of a society, that have here been suggested as underlying the 
customs of unilineal (patrilineal or matrilineal) succession are: 

1. The need for a formulation of rights over persons and 
things sufficiently precise in their general recognition as 
to avoid as far as possible unresolved conflicts. 

2. The need for continuity of the social structure as a system 
of relations between persons, such relations being de- 
finable in terms of rights and dutiesTI 

By American ethnologists who object to the method of 
explanation adopted in the preceding argument it is said that any 
sociological laws that can be formulated must necessarily be 
truisms. The laws formulated above, if they be true, as I believe, 
even if not adequately expressed, may be truisms. But even so 
they would seem to need to be brought to the attention of at least 
some ethnologists. A recent writer on the subject of matrilineal 
and patrilineal succession^ makes the following statements : 
'Unilateral institutions are in themselves anomalous and artificial. 
Matrilineal ones are doubly so.' 'Unilateral institutions, wherever 
found, represent deviations from the expectable, abnormalities 
in the social structure.' 'Unilateral reckoning contradicts the 
duality of parenthood and results in an unnatural stressing of one 

^ Ronald L. Olson, 'Clan and Moiety in North America', University of 
California Publications, Vol. 33, pp. 409, 411. 


side of the family to the exclusion of the other.' On the basis of 
these assertions he seems to conclude that unilineal determination 
of status must have had a single origin in some one aberrant people 
and to have spread from them, by a process of 'diffusion', to vast 
numbers of peoples in Europe, Asia, Africa, Austraha, Oceania 
and America. (One wonders, of course, why so many societies 
of so many different types should have accepted and retained 
such 'anomalous', 'abnormal' and 'unnatural' institutions.) 

I hope that the argument of this paper has shown, on the 
contrary, that unilineal institutions in some form, are almost, if 
not entirely, a necessity in any ordered social system. What is 
therefore unusual or rare (we need not say abnormal or anomalous 
and still less unnatural) is the discovery of a people such as the 
Teutonic peoples of Europe (apparently alone amongst Indo- 
European speaking peoples) maintaining for some period, until 
the coming of feudalism and Roman law, a system in which there 
is considerable, if not quite complete, avoidance of the unilineal 
principle, in which a person derives similar and equal rights 
through the father and through the mother.^ 

It might well be expected that such a paper as this would deal 
with the question of what general factors determine the selection 
by some people of the matrilineal and by others of the patrilineal 
principle in determining status or succession. My opinion is that 
our knowledge and understanding are not sufficient to permit us 
to deal with this problem in any satisfactory manner. 

^ There are systems of bilateral kinship with succession through both males 
and females in some parts of Indonesia, e.g. in the Ifugao of the Philippine 
Islands. The discussion of these would be complex and require space that is 
not available. 



FOR seventy-five years the subject of kinship has occupied 
a special and important position in social anthropology. 
I propose in this address to consider the methods that have 
been and are being used in that branch of our studies and the 
kinds of results that we may reasonably expect to arrive at by 
those methods/l' shall consider and compare two methods which 
I shall speak of a&4h^t of conjectural history and that of structural 
or sociological analysisTl 

One of these methods was first applied to some social in- 
stitutions by French and British (mostly Scots) writers of the 
eighteenth century. It was of this method that Dugald Stewart 
wrote in 1795: iTo this species of philosophical investigation, 
which has no appropriated name in our language, I shall take 
the liberty of giving the title of Theoretical or Conjectural History; 
an expression which coincides pretty nearly in its meaning with 
that of Natural History, as employed by Mr. Hume (see his 
Natural History of Religiofi), and with what some French writers 
have called Histoire Raisonnee.' I shall accept Dugald Stewart's 
suggestion and shall use the name 'conjectural history^' 

The method of conjectural history is used in a ijiimber of 
different ways. One is to attempt to base on general considerations, 
on what Dugald Stewart calls 'known principles of human nature', 
conjectures as to first beginnings — of political society (Hobbes), 
of language (Adam Smith), of religion (Tylor), of the family 
(Westermarck), and so on. Sometimes an attempt is made to deal 
with the whole course of development of human society, as in 
the works of Morgan, Father Schmidt and Elliot Smith. Some- 
times we are offered a conjectural history of the development 
of a particular institution, as in Robertson Smith's treatment of 
sacrifice. The special form of the method with which we shall 
be concerned in what follows is the attempt to explain a particular 
feature of one or more social systems by a hypothesis as to how 
it came into existence. 

^Presidential Address to the Royal Anthropological Institute, 1941. 
Reprinted from the Joiirnal of the Royal Anthropological Institute. 



An early example of the method of conjectural history applied 
to kinship is to be found in the essay on Primitive Marriage 
published by John F. M'Lennan in 1865. You will remember 
the two principal theses put forward in that book: the origin of 
the custom of exogamy from marriage by capture, and the pro- 
position that 'the most ancient system in which the idea of blood 
relationship was embodied was a system of kinship through 
females only'. Six years later there appeared The Systems of 
Consanguinity and Affinity of Lewis Morgan, a monument of 
scholarly, patient research in the collection of data, to be followed 
in 1877 by his Ancient Society, in which he offered a conjectural 
outline history of the whole course of social development. These 
works of M'Lennan and Morgan were followed by a considerable 
mass of literature, which has continued to be produced down to 
the present day, in which the method of conjectural history has 
been applied in different forms to various features of kinship 

\ As I think you know, I regard the pursuit of this method as one 
of the chief obstacles to the development of a scientific theory of 
human society. But my position has often been misunderstood. 
My objection to conjectural history is not that it is historical, 
but that it is conjectural. History shows us how certain events or 
changes in the past have led to certain other events or conditions, 
and thus reveals human life in a particular region of the world as a 
chain of connected happenings. But it can do this only when there 
is direct evidence for both the preceding and succeeding events 
or conditions and also some actual evidence of their interconnection. 
In conjectural history we have direct knowledge about a state of 
affairs existing at a certain time and place, without any adequate 
knowledge of the preceding conditions and events, about which 
we are therefore reduced to making conjectures. To establish any 
probability for such conjectures we should need to have a know- 
ledge of laws of social development which we certa inly do not 
possess and to which I do not think we shall ever attainTj 

My own study of kinship began in 1904 under "iGvers, when 
I was his first and at that time his only student in social anthropo- 
logy, having for three years previously studied psychology under 
him. I owe a great deal to that contact with Rivers, and more rather 
than less because from the outset it appeared that we disagreed 
on the subject of method. For Rivers followed the method of 


conjectural history, at first under the influence of Morgan, and 
later in the form of what he called ethnological analysis, as 
exemplified in his History of Melanesian Society (1914a). But in 
his field work Rivers had discovered and revealed to others the 
importance of the investigation of the behaviour of relatives to one 
another as a means of understanding a system of kinship. In 
what follows I shall be criticising one side of Rivers' work, but 
the position I now hold is the one I held in my friendly discussions 
with him during a period of ten years, ending in an agreement to 
go on disagreeing. My esteem for Rivers as man, as teacher, and 
as scientist, is in no way diminished by the fact that I find myself 
obliged to criticise adversely his use of the method of conjectural 

At the outset it is necessary to give a definition. I shall use the 
term 'kinship system' as short for a system of kinsliip and marriage 
or kinship and affinity. It is a pity that there is no inclusive term 
in English for all relationships which result from the existence 
of the family and marriage. It would be very tiresome to speak 
all the time of a system of kinship and affinity. I hope, therefore, 
that my use of the term will be accepted. It need not lead to 

The unit of structure from which a kinship system is built 
up is the group which I call an 'elementary family', consisting 
of a man and his wife and their child or children, whether they 
are living together or not. A childless married couple does not 
constitute a family in this sense. Children may be acquired, and 
thus made members of an elementary family, by adoption as well 
as by birth. We must also recognise the existence of compound 
families. In a polygynous family there is only one husband with 
two or more wives and their respective children. Another form 
of compound family is produced in monogamous societies by 
a second marriage, giving rise to what we call step -relationships 
and such relationships as that of half-brothers. Compound families 
can be regarded as formed of elementary families with a common 

The existence of the elementary family creates three special 
kinds of social relationship, that between parent and child, that 
between children of the same parents (siblings), and that between 
husband and wife as parents of the same child or children. A 
person is born or adopted into a family in which he or she is 


son or daughter and brother or sister. When a man marries and 
has children he now belongs to a second elementary family, in 
which he is husband and father. This interlocking of elementary 
families creates a network of what I shall call, for lack of any better 
term, genealogical relations, spreading out indefinitely. 

The three relationships that exist within the elementary 
family constitute what I call the first order. Relationships 
of the second order are those which depend on the connection 
of two elementary families through a common member, and are 
such as father's father, mother's brother, wife's sister, and so on. 
In the third order are such as father's brother's son and mother's 
brother's wife. Thus we can trace, if v/e have genealogical in- 
formation, relationships of the fourth, fifth or nth order. In any 
given society a certain number of these relationships are recog- 
nised for social purposes, i.e. they have attached to them certain 
rights and duties, or certain distinctive modes of behaviour. It is 
the relations that are recognised in this way that constitute what 
I am calling a kinship system, or, in full, a system of kinship and 

A most important character of a kinship system is its range. 
In a narrow range system, such as the English system of the 
present day, only a limited number of relatives are recognised 
as such in any way that entails any special behaviour or any 
specific rights and duties. In ancient times in England the range 
was wider, since a fifth cousin had a claim to a share of the wergild 
when a man was killed. In systems of very wide range, such as are 
found in some non-European societies, a man may recognise 
many hundreds of relatives, towards each of whom his behaviour 
is qualified by the existence of the relationship. 

It must be noted also that in some societies persons are re- 
garded as being connected by relationships of the same kind 
although no actual genealogical tie is known. Thus the members 
of a clan are regarded as being kinsmen, although for some of them 
it may not be possible to show their descent from a common 
ancestor. It is this that distinguishes what will here be called a 
clan from a lineage. 

Thus a kinship system, as I am using the term, or a system 
of kinship and affinity if you prefer so to call it, is in the first place 
a system of dyadic relations between person and person in a 
community, the behaviour of any two persons in any of these 


relations being regulated in some way, and to a greater or less 
extent, by social usage. 

A kinship system also includes the existence of definite social- 
groups. The first of these is the domestic family, which is a group 
of persons who at a particular time are living together in one 
dwelling, or collection of dwellings, with some sort of economic 
arrangement that we may call joint housekeeping. There are many 
varieties of the domestic family, varying in their form, their size, 
and the manner of their common life. A domestic family may 
consist of a single elementary family, or it may be a group in- 
cluding a hundred or more persons, such as the zadniga of the 
Southern Slavs or the taravad of the Nayar. Important in some 
societies is what may be called a local cluster of domestic families. 
In many kinship systems unilinear groups of kindred — lineage 
groups, clans and moieties — play an important part. 

By a kinship system, then, I mean a network of social relations 
of the kind just defined, which thus constitutes part of that total 
network of social relations that I call social structure. The rights 
and duties of relatives to one another and the social usages that 
they observe in their social contacts, since it is by these that the 
relations are described, are part of the system. I regard ancestor- 
worship, where it exists, as in a real sense part of the kinship 
system, constituted as it is by the relations of living persons to their 
deceased kindred, and afi^ecting as it does the relations of living 
persons to one another. The terms used in a society in addressing 
or referring to relatives are a part of the system, and so are the 
ideas that the people themselves have about kinship. 

You will perceive that by using the word 'system' I have made 
an assumption, an important and far-reaching assumption; for 
that word implies that whatever it is applied to is a complex 
unity, an organised whole^My explicit hypothesis is that between 
the various features of a particular kinship system there is a 
complex relation of interdependence. The formulation of this 
working hypothesis leads immediately to the method of socio- 
logical analysis, by which we seek to discover ^the nature of 
kinship systems as systems, if they be really such,,' For this pur- 
pose we need to make a systematic comparison of a sufficient 
number of sufficiently diverse systems.J^e must compare them, 
not in reference to single, superficial, and therefore immediately 
observable characters, but as wholes, as systems, and in reference, 


therefore, to general characters which are only discovered in the 
process of comparisoni Our purpose is to arrive at valid abstrac- 
tions or general ideas in terms of which the phenomena can be 
described and classified. 

I propose to illustrate the two methods, that of conjectural 
history and that of system, analysis, by means of a particular 
example, and for this purpose I select a peculiar feature of the 
kinship terminology of a number of scattered tribes. When Morgan 
made his study of the terminology of kinship in North American 
tribes, he noted certain peculiarities in the terms for cousins. 
In the Choctaw tribe he found that a man calls his father's sister's 
son by the same term of relationship that he applies to his own 
father and his father's brother. We may say that the father's 
sister's son is thus treated in the terminology as though he were 
a younger brother of the father. Reciprocally a man calls his 
mother's brother's son by the term for 'son'. Consistently with 
this he applies one term of relationship to his father's sister 
and her daughter, and speaks of his mother's brother's daughter 
as a 'daughter'. In the Omaha tribe, on the other hand, Morgan 
found that a man calls his mother's brother's son 'uncle', i.e. 
mother's brother, and calls his mother's brother's daughter 
'mother', so that reciprocally he speaks of his father's sister's son 
by the term that he uses for his sister's son, and a woman uses 
a single term for her own son, her sister's son and her father's 
sister's son. Figs, i and 2 will help to make these terminologies 

Terminologies similar to the Omaha are found in a number of 
regions: (i) in the Siouan tribes related to the Omaha, such as 
the Osage, Winnebago, etc.; (2) in certain Algonquian tribes, of 
which we may take the Fox Indians as an example; (3) in an 
area of California which includes the Miwok; (4) in some tribes of 
East Africa, both Bantu and non-Bantu, including the Nandi and 
the BaThonga; (5) amongst the Lhota Nagas of Assam; and 
(6) in some New Guinea tribes. Terminologies similar to the 
Choctaw are found: (i) in other south-eastern tribes of the United 
States, including the Cherokee; (2) in the Crow and Hidatsa 
tribes of the Plains area; (3) amongst the Hopi and some other 
Pueblo Indians; (4) in the Tlingit and Haida of the north-west 
coast of America; (5) in the Banks Islands in Melanesia; and (6) 
in one Twi-speaking community of West Africa. 


There are some who would regard this kind of terminology 
as 'contrary to common sense', but that means no more than that 
it is not in accordance with our modern European ideas of kinship 
and its terminology. It ought to be easy for any anthropologist to 
recognise that what is common sense in one society may be the 

= O 



"I r 

A = O 











Fig. I— c 






1 1 
A = 


A = 













A = 














Fig. 2 — 



= Father 

m = 



= Brother 

d = 




= Son 

s = 


opposite of common sense in another. The Choctaw and Omaha 
terminologies do call for some explanation; but so does the 
English terminology, in which we use the word 'cousin' for all 
children of both brothers and sisters of both mother and father — 
a procedure which would probably seem to some non-Europeans 
to be contrary not only to common sense but also to morals. What 
I wish to attempt, therefore, is to show you that the Choctaw 
and Omaha terminologies are just as reasonable and fitting in the 


social systems in which they occur as our o\\n terminology is 
in our own social system. 

I would point out that the Choctaw^ system and the Omaha 
system exhibit a single structural principle applied in different 
ways, in what we may perhaps call opposite directions. We shall 
therefore consider them together, as varieties of a single species. 

Attempts have been made to explain these terminologies by 
the method of conjectural history. The first was that of Kohler in 
1897, in his essay 'Zur Urgeschichte der Ehe', Kohler set out to 
defend Morgan's theory of group -marriage, and used the Choctaw 
and Omaha systems for his argument. He explained the Choctaw 
terminology as the result of marriage with the mother's brother's 
wife, and the Omaha system as the result of a custom of marriage 
w4th the wife's brother's daughter. Kohler's essay was reviewed 
by Durkheim (1898) in what was an important, if brief, con- 
tribution to the theory of kinship. He rejected Kohler's hypotheses, 
and pointed out the connection of the Choctaw and Omaha systems 
with matrilineal and patrilineal descent respectively. 

The subject was considered again by Rivers in reference to the 
Banks Islands, and, without bringing in, as Kohler had done, 
the question of group-marriage, he explained the Banks Islands 
terminology as resulting from a custom of marriage with the 
mother's brother's widow. Gifford (1916), having found the 
characteristic feature of the Omaha system in the Miwok of 
California, followed the lead of Kohler and Rivers, and explained 
it as the result of the custom of marriage with the wife's brother's 
daughter. About the same time, and independently, Mrs. Seligman 
(1917) offered the same explanation of the Omaha feature as it 
occurs in the Nandi and other tribes of Africa. 

Let me summarise the argument with reference to the Omaha 
type. The hypothesis is that in certain societies, mostly having 
a definite patrilineal organisation, a custom was for some reason 
adopted of permitting a man to marry his wife's brother's daughter. 
Referring to Fig, 3, this means that D would be allowed to 
marry f. When such a marriage occurred, then for G and h, f, 
who is their mother's brother's daughter, would become their 
step-mother, and E, their mother's brother's son, would become 
the brother of their step-mother. The hypothesis then assumes 
that the kinship terminology was so modified as to anticipate this 
form of marriage wherever it might occur. G and h will call f, 



their mother's brother's daughter and therefore their possible 
future step-mother, 'mother', and her brother E they will call 
'mother's brother'. Reciprocally f will call G 'son' and E will call 
him 'sister's son'. There is an exactly parallel argument for the 
Choctaw^ system. A custom arises by which a man may occasionally 
marry the widow of his mother's brother. In the figure, G would 
marry b, the wife of his mother's brother A. Thus E and f \vould 
become his step-children. If this marriage is anticipated in the 
terminology, then E and f will call G 'father' and h 'father's 

Let us note that in the Omaha tribe and in some others having 
a similar terminolog}' it is regarded as permissible for a man to 
marry his wife's brother's daughter. Marriage with the mother's 
brother's v.idow does not seem to occur regularly with the Choctaw 

A = 

= O 


O = 















c are 



and sist 


terminology, and does certainly occur without it, even in tribes 
with an Omaha terminology such as the BaThonga. 

The basis of what we may call the Kohler hypothesis is the 
obvious fact that in each of the two varieties the terminology 
and the special form of marriage are consistent; the two things 
fit together in what may be called a logical way. This, I think, 
anyone can see by inspection of the data. But the hypothesis goes 
far beyond this. It supposes that there is some sort of causal 
connection such that the marriage custom can be said to have 
caused, produced, or resulted in, the special terminology. No 
evidence is adduced that this is actually the way in \\hich things 
happened. The argument is entirely a priori. It is the essential 
weakness of conjectural history that its hypotheses cannot be 
verified. Thus this hypothesis cannot be considered as anything 
more than a speculation or conjecture as to how things might 
have happened. 


Now it would be equally plausible to suggest that the special 
form of marriage is the result of the terminology. If, as in the 
terminology of the Omaha type, I treat my wife's brother's 
daughter as being the younger sister of my wife, and, by the 
custom of the sororate, it is considered proper for me to marry my 
wife's younger sister, then I might well be permitted to marry the 
woman who, in the terminological system, is treated as such, 
namely her brother's daughter. This hypothesis is, of course, 
equally lacking in proof. If we adopt the Kohler hypothesis the 
terminology is conceived to be in some sense explained, but there 
is no explanation of the marriage custom. By the alternative 
hypothesis the marriage custom is explained, but the terminology 
is not. I do not see how there can be any ground for a choice 
of one of these two hypotheses in preference to the other except 
purely personal predilection. 

However, while we could conceive of the marriage custom 
as being the immediate result of the terminology in a society 
which already has sororal polygyny, the terminology cannot be 
the immediate result of the marriage custom without the con- 
comitant action of some other undetermined factor. We have ex- 
amples of societies in which a man sometimes marries the widow of 
his mother's brother, but only uses the terminology which this 
marriage makes appropriate after the marriage has taken place. 
Although we have no recorded instance of this procedure in 
marriage with the wife's brother's daughter it is at least conceivable 
that it might occur. What is lacking in the hypothesis we are 
examining is some reason why the whole terminology should be 
adjusted so as to fit a particular form of marriage which only 
occasionally occurs. 

Let us now leave the hypothesis and examine the structural 
principles of those kinship systems in which this terminology 
occurs, whether in the Choctaw or the Omaha form. It is necessary, 
however, to say something on the subject of kinship terminologies, 
about which there has been a great deal of controversy. Morgan's 
first interest in the subject was as an ethnologist, i.e. one seeking 
to discover the historical relations of the peoples of the earth. He 
thought that by collecting a sufficient sample of terminologies 
and comparing them he could reveal the historical relation of the 
American Indians (the Ganowanian peoples as he called them) 
to the peoples of Asia. In the course of his work, however, he 


decided that these terminologies could be used to infer the former 
existence of forms of social organisation. He supposed that the 
classificatory terminology which he found in North American 
tribes such as the Iroquois was inconsistent with the form of 
social organisation with which it is actually found, and therefore 
could not have arisen in a society so organised, but must be a 
'survival' from some different kind of social system. 

This was, of course, pure assumption, but it is the kind of 
assumption that the method of conjectural history encourages 
us to make, often unconsciously or implicitly. Morgan was thus 
led to a hypothesis that is one of the most fantastic in a subject 
that is full of fantastic hypotheses. The truth is that he had quite 
failed to understand the nature and function of the classificatory 
terminology. There is nothing that so effectively prevents the 
perception and understanding of things as they are as hypotheses 
of conjectural history, or the desire to invent such hypotheses. 

One of Morgan's early critics, Starcke (1889), was, I believe, 
the first to maintain the position which has always been my own. 
He held that in general a kinship nomenclature is 'the faithful 
reflection of the juridical relations which arise between the nearest - 
kinsfolk in each tribe'. He condemned as unsound the attempt 
to use such nomenclatures to make historical reconstructions of 
past societies. It would be interesting to consider why it is that 
Starcke has had so few followers and Morgan so many, but that 
I cannot here undertake. 

In 1909 Kroeber published in our Journal a paper on 'Classi- 
ficatory Systems of Relationship '.To the contentions of that paper 
Rivers made a reply in his lectures on Kinship and Social Or- 
ganisation (1914&), and Kroeber answered the criticisms of 
Rivers in his California Kinship Systems (1917). 

I discussed Kroeber's paper with Rivers when it appeared 
and found myself in the position of disagreeing with both sides 
of the controversy. Kroeber wrote: 'Nothing is more precarious 
than the common method of deducing the recent existence of 
social or marital institutions from a designation of relationship.' 
This is a restatement of Starcke's contention of 1889, and with 
it I was, and still am, in complete agreement, thereby disagreeing 
with Rivers. Kroeber also wrote: 'It has been an unfortunate 
characteristic of the anthropology of recent years to seek in a 
great measure specific causes for specific events, connection 


between which can be established only through evidence that 
is subjectively selected. On wider knowledge and freedom from 
motive it is becoming increasingly apparent that causal explanations 
of detached anthropological phenomena can be but rarely found 
in other detached phenomena.' With this statement I am in 

But both Kroeber and Rivers seemed to agree that causal 
explanations are necessary for the constitution of what Kroeber 
calls 'true science'. For Rivers anthropology is a true science 
because, or to the extent that, it can show causal connections; 
for Kroeber it is not a true science. Here I disagree with both 
Kroeber and Rivers, holding that a pure theoretical science 
(whether physical, biological or social) is not concerned with 
causal relations in this sense. The concept of cause and effect 
belongs properly to applied science, to practical life and its 
arts and techniques and to history. 

This brings us to the crux of the Rivers-Kroeber debate. 
Rivers held that the characteristics of a kinship nomenclature 
are determined by social or sociological factors, that particular 
features of terminology result from particular features of social 
organisation. Against this Kroeber held that the features of a 
system of terminology 'are determined primarily by language' 
and 'reflect psychology not sociology'. 'Terms of relationship', 
he wrote, 'are determined primarily by linguistic factors and 
are only occasionally, and then indirectly, affected by social 
circumstances.' But in his later paper Kroeber explains that 
what he calls psychological factors 'are social or cultural 
phenomena as thoroughly and completely as institutions, beliefs 
or industries are social phenomena'. His thesis is therefore con- 
cerned with a distinction between two kinds of social phenomena. 
One of these he calls institutional, defined as 'practices connected 
with marriage, descent, personal relations, and the like'. These 
are what he called in his first paper 'social factors'. The other 
kind he speaks of as the 'psyche' of a culture, 'that is, the ways of 
thinking and feeling characteristic of the culture'. These constitute 
what he calls the psychological factors. 

Thus Kroeber's thesis, on its positive side, is that similarities 
and differences of kinship nomenclature are to be interpreted or 
understood by reference to similarities and differences in the 
general 'manner of thought'. On its negative side, and it is with 


this that we are concerned, Kroeber's thesis is that there is no 
regular close connection between similarities and differences 
of kinship nomenclature and similarities and differences of 
'institutions', i.e. practices connected with marriage, descent and 
personal relations. He admits, in 19 17, the existence of 'undoubted 
correspondence of terminology and social practice in certain 
parts of Australia and Oceania', but denies that such are to be 
found in California. It may be pointed out that in Australia and 
Oceania they have been deliberately looked for, in California they 
have not. It may well be that in the remnants of Californian 
tribesJt is now too late to look for them. 

\ln_Qpposition to Kroeber, and in a certain sense in agreement 
with Rivers, i I hold that all over the world there are impor- 
tant correspondences between kinship nomenclature and social 
practices. Such correspondences are not to be simply assumed; 
they must be demonstrated by field ^^ork and comparative 
analysis^ But their absence may not be assumed either; and 
Kroeber's arguments from their alleged absence in California 
remain, I think, entirely unconvincing.^ 

For Kroeber the kinship nomenclature of a people represents 
their general manner of thought as it is applied to kinship. But 
the institutions of a people also represent their general manner of 
thought about kinship and marriage. Are we to suppose that in 
Californian tribes the way of thinking about kinship as it appears 
on the one hand in the terminology and on the other hand in 
social customs are not merely different but are not connected? 
Tliis seems to be in effect what Kroeber is proposing. 

Kroeber pointed out in 19 17 that his original paper represented 
'a genuine attempt to understand kinship systems as kinship 
systems'. But by 'kinship system' Kroeber means only a system 
of nomenclature. Moreover, Kroeber is an ethnologist, not a 
social anthropologist. His chief, if not his sole, interest in the 
subject is in the possibility of discovering and defining the 
historical relations of peoples by comparison of their systems of 

My own conception is that the nomenclature of kinship is 
an intrinsic part of a kinship system, just as it is also, of course, 
an intrinsic part of a language. The relations between the nomen- 
clature and the rest of the system are relations within an ordered 
whole. My concern, both in field work in various parts of the world 


and in comparative studies, has been to discover the nature of 
these relations. 

In the actual study of a kinship system the nomenclature is 
of the utmost importance. It affords the best possible approach 
to the investigation and analysis of the kinship system as a whole. 
This, of course, it could not do if there were no real relations of 
interdependence between the terminology and the rest of the 
system. That there are such relations I can affirm from my own 
field work in more than one region. It will be borne out, I believe, 
by any anthropologist who has made a thorough field study 
of a kinship system.^ 

I have dealt with the controversy between Kroeber and 
Rivers because, as both the controversialists point out, the real 
issue is not simply one concerning kinship terms, but is a very 
important question of the general method of anthropological 
studies. It seemed to me that I could best make clear my own 
position by showing you how it differs from that of Rivers on 
the one side and that of Kroeber on the other. 

Kinship systems are made and re-made by man, in the same 
sense that languages are made and re-made, which does not mean 
that they are normally constructed or changed by a process of 
deliberation and under control of conscious purpose. A language 
has to work, i.e. it has to provide a more or less adequate in- 
strument for communication, and in order that it may work it 
has to conform to certain general necessary conditions. A morpho- 
logical comparison of languages shows us the different ways in 
which these conditions have been complied with by using different 
morphological principles such as inflection, agglutination, word 
order, internal modification or the use of tone or stress. A kinship 
system also has to work if it is to exist or persist. It has to provide 
an orderly and workable system of social relations defined by social 
usage. A comparison of different systems shows us how workable 
kinship systems have been created by utilising certain structural 
principles and certain mechanisms. .^ 

One common feature of kinship systems is the recognition 
of certain categories or kinds into which the various relatives of a 

^ My position has been misunderstood and consequently misrepresented 
by Dr. Opler (1937ft) in his paper on ' Apache Data concerning the Relation of 
Kinship Terminology to Social Classification'; but the first two paragraphs 
of another of Dr. Opler's papers (1937a) on 'Chiricahua Apache Social Or- 
ganisation', state what was at that time his, and is also my, point of view. 


single person can be grouped. The actual social relation between a 
person and his relative, as defined by rights and duties or socially 
approved attitudes and modes of behaviour, is then to a greater 
or less extent fixed by the category to which the relative belongs. 
The nomenclature of kinship is commonly used as a means of 
establishing and recognising these categories, A single term may 
be used to refer to a category of relatives and different categories 
will be distinguished by different terms. 

Let us consider a simple example from our own system. 
We do what is rather unusual in the general nm of kinship 
systems: we regard the father's brother and the mother's brother 
as relatives of the same kind of category. We apply a single term 
originally denoting the mother's brother (from the Latin avun- 
culus) to both of them. The legal relationship in English law, 
except for entailed estates and titles of nobility, is the same for a 
nephew and either of his uncles; for example, the nephew has the 
same rights of inheritance in case of intestacy over the estate of 
either. In what may be called the socially standardised behaviour 
of England it is not possible to note any regular distinction made 
between the maternal and the paternal uncle. Reciprocally the 
relation of a man to his different kinds of nephews is in general the 
same. By extension, no significant difference is made between 
the son of one's mother's brother and the son of one's father's 
brother. ^J^alJ?^^ 

In Montenegro, on the contrary, to take another European 
system, the father's brothers constitute one category and the 
mother's brothers another. These relatives are distinguished by 
different terms, and so are their respective wives, and the social 
relations in which a man stands to his two kinds of uncles show 
marked differences. 

There is nothing 'natural' about the English attitude towards 
uncles. Indeed many peoples in many parts of the world would 
regard this failure to distinguish between relatives on the father's 
side and those on the mother's side as unnatural and even im- 
proper. But the terminology is consistent with our whole kinship 

The kinship systems with which we shall be concerned here 
all have certain forms of what Morgan called the 'classificatory' 
terminology. What Morgan meant by this term is quite clear from 
his writings, but his definition is often ignored, perhaps because 


people do not bother to read him. A nomenclaturejs-classiiicatory 
when it uses terms wliich primariFy appIy~toTineal relatives, such 
as 'father', to refer also to collateral relatives. Thus by Morgan's 
definition the English word 'uncle' is not a classificatory term, 
but the very opposite, since it is used only for collateral relatives. 
Kroeber (1909) criticises Morgan and rejects his conception of 
classificatory terminologies, and then proceeds to make use of the 
same distinction by taking as one of the important features of 
terminologies the extent to which they separate or distinguish 
lineal from collateral relatives. It seems to be merely the word 
'classificatory' that Kroeber does not like. Doubtless it is not the 
ideal word; but it has long been in use and no better one has been 
suggested, though others have been put forward. 

I do not propose to deal with all systems in which the classi- 
ficatory principle is applied in the terminology, but only with a 
certain widespread type. In these systems the distinction between 
lineal and collateral relatives is clearly recognised and is of great 
importance in social life, but it is in certain respects subordinated 
to another structural principle, which can be spoken of as the 
principle of the solidarity of the sibling group. A group of siblings 
is constituted by the sons and daughters of a man and his wife in 
monogamous societies, or of a man and his wives where there 
is polygyny, or of a woman and her husbands in polyandrous 
communities. The bond uniting brothers and sisters together 
into a social group is everywhere regarded as important, but it is 
more emphasised in some societies than in others. The solidarity 
of the sibling group is shown in the first instance in the social 
relations between its members. 

From this principle there is derived a further principle which 
I shall speak of as that of the unity of the sibling group. This 
refers not to the internal unity of the group as exhibited in the 
behaviour of members to one another, but to its unity in relation 
to a person outside it and connected with it by a specific relation 
to one of its members. 

A diagram may help the discussion. Fig. 4 represents a sibling 
group of three brothers and two sisters, to which Ego is related 
by the fact that he is the son of one of the three men. In the 
kinship systems with which I am now dealing. Ego regards himself 
as standing in the same general kind of relation to all the members 
of the group. For him it constitutes a unity. His relation to the 


brothers and sisters of his father is conceived as being of the same 
general kind as his relation to his father. Within the group, 
however, there are two principles of differentiation, sex and 
seniority, which have to be taken into account. In systems in which 
seniority is not emphasised a man treats his father's brothers, 
both older and younger, as being like his father. He refers to 
them or addresses them by the same term of kinship that he applies 
to his own father, and in certain important respects his behaviour 
towards them is similar to his behaviour towards his own father. 
What defines this behaviour is, of course, different in different 
systems. Where seniority is strongly emphasised, a man may 
distinguish betv^een the senior brother and the junior brother 
either in behaviour alone or both in behaviour and terminology, 

A A A 


Fig. 4 

but there still remains a common element in the pattern of be- 
haviour towards all 'fathers'. 

The difference of sex is more important than the difference of 
seniority, and in this matter there is considerable variation in the 
systems we are considering. But in quite a considerable number 
of systems, in different parts of the world, there are certain features 
of a man's relationship to his father's sister which can be correctly 
described by saying that he regards her as a sort of female father. 
In some of these systems he actually calls her 'female father', 
or some modification of the term for father. If it seems to you 
impossible that a man should regard his father's sister as a relative 
of the same kind as his own father, this is because you are thinking, 
not about social relationships as defined by modes of behaviour, 
with which we are here concerned, but about the physiological 
relationship, which is irrelevant. 

The same kind of thing happens with the sibling group of the 
mother. The mother's sisters are treated as relatives of the same 
kind as the mother, both in terminology and in certain principles 
of behaviour or attitude. In a number of systems the mother's 
brother is also treated as a relative of the same kind as the mother. 


He may be called 'male mother', as in Bantu tribes of Africa and in 
Tonga in the Pacific. If the principle of seniority is stressed, the 
mother's brothers may be distinguished according as they are 
older or younger than the mother. 

Those of you who have never had any direct contact with 
systems of this kind may find it difficult to comprehend how a 
father's sister can be regarded as a female father or a mother's 
brother as a male mother. This is due to the difficulty of dissociating 
the terms 'father' and 'mother' from the connotations they have in 
our own social system. It is absolutely essential to do this if the 
kinship systems of other societies are ever to be understood. 
Perhaps it will help somewhat if I refer to another terminology 
which seems to us peculiar. Most of the systems with which I am 
now dealing have a word for 'child', or words for 'son' and 
'daughter', which a man applies to his own children and his 
brother's children, and a woman applies to her own children and 
her sister's children. But in some Australian tribes there are two 
different words for 'child'. One is used by a man for his own child 
(or his brother's child) and by a woman for her brother's child; 
the other is used by a woman for her own or her sister's child, and 
by a man for his sister's child. I think you will see that this is 
another way of expressing in the terminology the unity that links 
brother and sister in relation to the child of either of them. I am 
called by one term by my father and his brothers and sisters; 
and by another term by my mother and her sisters and brothers. 

The same principle, that of the unity of the sibling group, 
is applied to other sibling groups. Thus the father's father's 
brother is regarded as belonging to the same category as the 
father's father, with the result that his son is a somewhat more 
distant relative of the same kind as the father and his brothers. 
By means of such extension of the basic principle, a very large 
number of collateral relatives of different degrees of distance can 
be brought under a limited number of categories. A man may 
have many, even hundreds, of relatives whom he thus classifies 
as 'fathers', 'brothers', 'mother's brothers' and so on. But there are 
different ways in which this extension of the basic classificatory 
principle can be applied, so that there result systems of different 
types. What is conmion to them all is that they make some use 
of this structural principle which I have briefly illustrated. 

What I am trying to show-you is that the classificatory 


terminology_iA_Ji_jnethod of providing a wide-riange Jciii^ipL^ 
organisation, by making use of the unity of the sibling group 
in order to establish a few categories of relationship under which 
a very large number of near and distant relatives can be included.. 
For all the relatives who are denoted by one term, there is normally ( 
some element of attitude or behaviour that is regarded as ap- ' 
propriate to them and not to others. But within a category there 
may be and always are important distinctions. There is, first, 
the very important distinction between one's own father and 
his brother. There are distinctions within the category between 
nearer and more distant relatives. There is sometimes an im- 
portant distinction between relatives of a certain category who 
belong to other clans. There are other distinctions that are made 
in different particular systems. Thus the categories represented 
by the terminology never give us anything more than the skeleton 
of the real ordering of relatives in the social life. But in every 
system that I have been able to study they do give us this skeleton. 

If this thesis is true, if this is what the classificatory ter- 
minology actually is in the tribes in which it exists, it is obvious 
that Morgan's whole theor\' is entirely ungrounded. The classi- 
ficatory system, as thus interpreted, depends upon the recognition 
of the strong social ties that unite brothers and sisters of the same 
elementary family, and the utilisation of this tie to build up a 
complex orderly arrangement of social relations amongst kin. 
It could not come into existence except in a society based on 
the elementary family. Nowhere in the world are the ties between 
a man and his own children or between children of one father 
stronger than in Australian tribes, which, as you know, present 
an extreme example of the classificatory terminology. 

The internal solidarity of the sibling group, and its unity 
in relation to persons connected with it, appear in a great number 
of different forms in different societies. I cannot make any attempt 
to deal with these, but for the sake of the later arginnent I will 
point out that it is in the light of this structural principle that we 
must interpret the customs of sororal polygyny (marriage with 
two or more sisters), the sororate (marriage with the deceased 
wife's sister), adelphic polyandry (marriage of a woman with two 
or more brothers, by far the commonest form of polyandry), and 
the levirate (marriage with the brother's widow). Sapir, using the 
method of conjectural history, has suggested that the classificatory 


terminology may be the result of the customs of the levirate and 
sororate. That the two things are connected is, I think, clear, 
but for the supposed causal connection there is no evidence w^hat- 
ever. Their real connection is that they are different ways of 
applying or using the principle of the unity of the sibling group, 
and they may therefore exist together or separately. 

An organisation into clans or moieties is also based on the 
principle of the solidarity and unity of the sibling group in com- 
bination with other principles. Tylor suggested a connection 
between exogamous clans and the classificatory terminology. 
Rivers put this in terms of conjectural history, and argued that the 
classificatory terminology must have had its origin in the or- 
ganisation of society into exogamous moieties. 


It is necessary, for our analysis, to consider briefly another 
aspect of the structure of kinship systems, namely the division 
into generations. The distinction of generation has its basis in the 
elementary family, in the relation of parents and children. A 
certain generalising tendency is discoverable in many kinship 
systems in the behaviour of relatives of different generations. 
Thus we find very frequently that a person is expected to adopt 
an attitude of more or less marked respect towards all his relatives 
of the first ascending generation. There are restraints on behaviour 
which maintain a certain distance or prevent too close an in- 
timacy. There is, in fact, a generalised relation of ascendancy 
and subordination between the two generations. This is usually 
accompanied by a relation of friendly equality between a person 
and his relatives of the second ascending generation. The nomen- 
clature for grandparents and grandchildren is of significance in 
this connection. In some classificatory systems, such as those of 
Australian tribes, the grandparents on the father's side are dis- 
tinguished, in terminology and in behaviour, from those on the 
mother's side. But in many classificatory systems the generalising 
tendency results in all relatives of the generation being classed 
together as 'grandfathers' and 'grandmothers'. 

We may note in passing that in classificatory terminologies 
of what Morgan called the Malayan type and Rivers the Hawaiian 
type, this generalising process is applied to other generations, 


SO that all relatives of the parents' generation may be called 
'father' and 'mother' and all those of one's own generation may 
be called 'brother' and 'sister'. 

There are many kinship systems in various parts of the world 
that exhibit a structural principle which I shall speak of as the 
combination of alternate generations. This means that relatives of 
the grandfather's generation are thought of as combined with those 
of one's own generation over against the relatives of the parents' 
generation. The extreme development of this principle is to be 
seen in Australian tribes. I shall refer to this later. 

While some systems emphasise the distinction of generations 
in their terminology or in their social structure, there are also 
systems in which relatives of two or more generations are in- 
cluded in a single category. So far as I have been able to make a 
comparative study, the various instances of this seem to fall into 
four classes. 

In one class of instances the term of relationship does not 
carry a connotation referring to any particular generation and is 
used to mark off a sort of marginal region between non-relatives 
and those close relatives towards whom specific duties and over 
whom specific rights are recognised. The application of the term 
generally only implies that since the other person is recognised as a 
relative he or she must be treated with a certain general attitude 
of friendliness and not as a stranger. A good example is provided 
by the terms ol-le-sotwa and en-e-sotwa in Masai. I would include 
the English word 'cousin' in this class. 

A second class of instances includes those in which there is 
conflict or inconsistency between the required attitude towards a 
particular relative and the required general attitude towards the 
generation to which he belongs. Thus in some tribes in South- 
East Africa there is conflict between the general rule that relatives 
of the first ascending generation are to be treated with marked 
respect and the custom of privileged disrespect towards the 
mother's brother. This is resolved by placing the mother's 
brother in the second ascending generation and calling him 
'grandfather'. An opposite example is found in the Masai. A 
man is on terms of familiarity with all his relatives of the second 
descending generation, who are his 'grandchildren'. But it is felt 
that the relation between a man and the wife of his son's son 
should be one not of familiarity but of marked reserve. The 


inconsistency is resolved by a sort of legal fiction by which she is 
moved out of her generation and is called 'son's wife'. 

A third class of instances are those resulting from the struc- 
tural principle, already mentioned, whereby alternate generations 
are combined. Thus the father's father may be called 'older brother' 
and treated as such, and the son's son may be called 'younger 
brother'. Or a man and his son's son may be both included in a 
single category of relationship. There are many illustrations 
of this in Australian tribes and some elsewhere. An example 
from the Hopi will be given later. 

The fourth class of instances includes the systems of Choctaw 
and Omaha type and also certain others, and in these the dis- 
tinction betsveen generations is set aside in favour of another 
principle, that of the unity of the lineage group. 

Since the word lineage is often loosely used, I must explain 
what I mean by it. A patrilineal or agnatic hneage consists of a 
man and all his descendants through males for a determinate 
number of generations. Thus a minimal lineage includes three 
generations, and we can have lineages of four, five or n generations. 
A matrilineal lineage consists of a woman and all her descendants 
through females for a determinate number of generations. A 
lineage group consists of all the members of a lineage who are alive 
at a particular time. A clan, as I shall use the term here, is a 
group which, though not actually or demonstrably (by genealogies) 
a lineage, is regarded as being in some ways similar to a lineage. 
It normally consists of a number of actual lineages. Lineages, 
both patrilineal and matrilineal, exist implicitly in any kinship 
system, but it is only in some systems that the solidarity of the 
lineage group is an important feature in the social structure. 

Where lineage groups are important we can speak of the 
solidarity of the group, which shows itself in the first instance 
in the internal relations between the members. By the principle 
of the unity of the lineage group I mean that for a person who 
does not belong to the lineage but is connected with it through 
some important bond of kinship or by marriage, its members 
constitute a single category, with a distinction within the category 
between males and females, and possibly other distinctions also. 
When this principle is applied in the terminology a person 
connected with a lineage from outside applies to its members, of 
one sex, through at least three generations, the same term of 


relationship. In its extreme development, as applied to the clan, a 
person connected with a clan in a certain way applies a single 
term of relationship to all members of the clan. An example will 
be given later. 

The Omaha type of terminology may be illustrated by the 
system of the Fox Indians, which has been carefully studied by 
Dr. Sol Tax (1937). The features of the system that are relevant 


A = 

gm ' 







A = 



fs [ 






A = 






sis [ 







Fig. s— Fox 

Father's Lineage 





to the argument are illustrated in the accompanying diagrams 
(Figs. 5-9).i 

In his own patrilineal lineage a man distinguishes his relatives 
according to generation as 'grandfather' (GF), 'father' (F), 
'older or younger brother' (B), 'son' (S), 'grandmother' (gm), 

^ In these diagrams A represents a male person and O a female. The sign 
= connects a man and his wife and the lines descending from it indicate their 
children. The letters (capitals for males and lower case for females) stand for the 
kinship terms of a classificatory system, in which the same term is applied to a 
number of relatives. GF stands for the term used in referring to a grandfather, 
and similarly gm for grandmother; the others are F, father, m, mother, ms, 
mother's sister, fs, father's sister, MB, mother's brother, FL, father-in-law, 
ml, mother-in-law, B, brother, sis, sister, BL, brother-in-law, si, sister-in-law, 
S, son, d, daughter, N, nephew (strictly speaking sister's son) n, niece (sister's 
daughter of a male) GC or gc, grandchild. 


'father's sister' (fs), 'sister' (sis), and 'daughter' (d). I would draw 
your attention to the fact that he appHes a single term, 'brother-in- 
law' (BL), irrespective of generation, to the husbands of the women 
of the lineage through three generations (his own and the two 
ascending generations), and that he calls the children of all these 
women by the same terms, 'nephew' (N) and 'niece' (n). Thus the 
women of Ego's own lineage of these generations constitute a sort 
of group, and Ego regards himself as standing in the same relation- 
ship to the children and husbands of all of them, although these 
persons belong to a number of different lineages. 


A = 

= O 

o = 

gm 1^ 






A = 

= O 




O = 






A = 

= O 



O = 
ms 1 








A = 

= O 

Fig. 6- 

M other's 


O = 








Turning to the mother's patrilineal lineage, it can be seen 
that a man calls his mother's father 'grandfather', but calls all the 
males of the lineage in the three succeeding generations 'mother's 
brother' (MB). Similarly he calls the women of these three 
generations, except his own mother, by a term translated as 
'mother's sister' (ms). He applies the term 'father' (F) to the 
husbands of all the women of the lineage through four generations 
(including the husband of the mother's father's sister) and the 
children of all these women are his 'brothers' and 'sisters'. 
He is the son of one particular woman of a unified group, 


and the sons of the other women of the group are therefore his 

A = 

A = 


O = 


Fig. 7 — Fox 

Father's Mother's Lineage 


A = O 
FL ml 

A = 

A = 



Fig. 8— Fox 

Wife's Lineage 

O = A 
Wife EGO 

In his father's mother's hneage Ego calls all the men and 
women throughout three generations 'grandfather' and 'grand- 
mother'. The children of these 'grandmothers' are all his 'fathers' 
and 'father's sisters', irrespective of generation. In his mother's 



mother's lineage he also calls all the males 'grandfather' and the 
females 'grandmother', but I have not thought it necessary to 
include a figure to show this. 

In his wife's lineage a man calls his wife's father by a term 
which we will translate 'father-in-law' (FL). It is a modification 
of the word for 'grandfather'.^ The sons and brother's sons of 
the 'fathers-in-law' are 'brothers-in-law' (BL), and the daughters 
are 'sisters-in-law' (si). The children of a 'brother-in-law* are 
again 'brother-in-law' and 'sister-in-law'. Thus these two terms 
are applied to the men and women of a lineage through three 


O - A 
ml I FL 

A = 


O = A 
Wife EGO 

Fig. 9 — Fox 
Wife's Mother's Lineage 

generations. The children of all these 'sisters-in-law' are 'sons' 
and 'daughters'. 

Fig. 9 shows the lineage of the wife's mother. In this lineage, 
through three generations, all the men are called 'father-in-law' 
and all the women 'mother-in-law'. 

Is the classification of relatives in the Fox terminology 
simply a matter of language, as some would have us believe? 
Dr. Tax's observations (1937) enable us to affirm that it is not. 
He writes: 

The kinship terminology is applied to all known relatives (even in 
some cases where the genealogical relationship is not traceable) so that 
the entire tribe is divided into a small number of types of relationship 

^ The Fox terms for father-in-law and mother-in-law are modifications 
of the terms for grandfather and grandmother. In the Omaha tribe the terms 
for grandparents, without modification, are applied to the parents-in-law and 
to those who are called 'father-in-law' and 'mother-in-law' in the Fox tribe. 


pairs. Each of these types carries with it a more or less distinct tradi- 
tional pattern of behaviour. Generally speaking, the behaviour of close 
relatives follows the pattern in its greatest intensity, that of farther 
relatives in lesser degree ; but there are numerous cases where, for 
some reason, a pair of close relatives ' do not behave towards each other 
at all as they should'. 

Dr. Tax goes on to define the patterns of behaviour for the 
various types of relationship. Thus the classification of relatives 
into categories, carried out by means of the nomenclature, or 
therein expressed, appears also in the regulation of social 
behaviour. There is good evidence that this is true of other 
systems of Omaha type, and, contrary to Kroeber's thesis, we may 
justifiably accept the hypothesis that it is probably true of all. 

Charts similar to those given here for the Fox Indians can 
be made for other systems of the Omaha type. I think that a 
careful examination and comparison of the various systems shows 
that, while there are variations, there is a single structural principle 
underlying both the terminology and the associated social structure. 
A lineage of three (or sometimes more) generations is regarded as a 
unity. A person is related to certain lineages at particular points: 
in the Fox tribe to the lineages of his mother, his father's mother, 
his mother's mother, his wife, and his wife's mother. In each 
instance he regards himself as related to the succeeding generations 
of the lineage in the same way as he is related to the generation 
with which he is actually connected. Thus all the men of his 
mother's lineage are his 'mother's brothers', those of his 'grand- 
mother's lineage his 'grandfathers', and those of his wife's lineage 
are his 'brothers-in-law'. 

This structural principle of the unity of the patrilineal 
lineage is not a hypothetical cause of the terminology. It is a 
principle that is directly discoverable by comparative analysis 
of systems of this type; or, in other words, it is an immediate 
abstraction from observed facts. 

Let us now examine a society in which the principle of the 
unity of the lineage group is applied to matrilineal lineages. 
For this I select the system of the Hopi Indians, which has been 
analysed in a masterly manner by Dr. Fred Eggan (1933) in a 
Ph.D. thesis which has, unfortunately, not yet been published.^ 

^ The thesis, in a revised form, has now been published. Eggan: Social 
Organisation of the Western Pueblos. The University of Chicago Press, 1950. 


The most significant features of the system are illustrated in the 
accompanying figures. 

A = 



A = 

B i si 


A ' 





A = 
MB [si 

1 ^ 


A = 






A = 
B 1 si 

1 1 



A = 



A = 

N L 

1 ci 


A = 


Fig. 10- 




1 1 


A man's own lineage is, of course, that of his mother. He 
distinguishes the women of his lineage by generation as 'grand- 
mother' (gm), 'mother' (m), 'sister' (sis), 'niece' (n), and 'grand- 
child' (gc). Amongst the men of his lineage he distinguishes his 
'mother's brothers' (MB), 'brothers' (B) and 'nephews' (N). 
But he includes his mother's mother's brother and his sister's 
daughter's son in the same category as his brothers. The structural 
principle exhibited here is that already referred to as the com- 
bination of alternate generations. It should be noted that a man 
includes the children of all men of his own lineage, irrespective 
of generation, in the same category as his own children. Fig. 10 
should be carefully compared with Fig. 5, for the Fox Indians, as 
the comparison is illuminating. 



In his father's Uneage a man calls all the male members 
through five generations 'father' and, with the exception of his 
father's mother (his 'grandmother'), he calls all the women 
'father's sister'. The husband of any woman of the lineage is a 
'grandfather', and the wife of any man of the lineage is a 'mother'. 

A = O 
GF gm 

A = O 
GF fa 

A = O 
GF fs 


A = O 
GF fs 

Fig. II — Hopi 
Father's Lineage 

The children of his 'fathers' are 'brothers' and 'sisters'. Fig. ii 
should be carefully compared with Fig. 6. 

In his mother's father's lineage a man calls all the men 
and women through four generations 'grandfather' and 'grand- 

The Hopi do not regard a man as related to his father's father's 
lineage as a whole, and the principle is therefore not applied to it. 
He does call his own father's father 'grandfather'. 

Dr. Eggan has shown that for the Hopi this classification 
of relatives into categories is not simply a matter of terminology 


or language, but is the basis of much of the regulation of 
social life. 

What is, I think, clearly brought out by a comparison of the 
Fox and Hopi systems is their fundamental similarity. By the 
theories of conjectural history this similarity is the accidental 
result of different historical processes. By my theory it is the 
result of the systematic application of the same structural principle, 
in one instance to patrilineal and in the other to matrilineal lineages. 

A = 


A = 
GF Igm 


A = 


A - 
GF gm 


A = 


= C 

A = 



A = 
GF gm 


A = 




A = 
GF gm 

12 — Hopi 
Father's Lineage 

I cannot, of course, discuss all the various systems of Choctaw 
and Omaha type. The variations that they show in certain features 
are very interesting and important. If you wish to test my theory 
you will examine them, or some of them, for yourselves, and the 
easiest way to analyse any system is to reduce it to a set of lineage 
charts similar to those given here for the Fox and the Hopi. 
For any system such a set of charts will reveal the exact way in 
which the general principle of the unity of the lineage is applied. 
The manner of application varies somewhat, but the principle 
appears in each system of the type. 

You will doubtless already have noticed that in these systems 
there are an extraordinary number of relatives of all ages to whom 


a man applies the terms 'grandfather' and 'grandmother'. There 
is, I beheve, a good reason for this, which should be briefly 
indicated. It is a general rule in societies having a classificatory 
terminology that for all the various relatives included under a 
single term there is some more or less definite pattern of behaviour 
which is regarded as normal or appropriate. But there are important 
differences in this matter. In certain instances the pattern can 
be defined by reference to specific rights and duties, or by specific 
modes of behaviour. For example, in the Kariera tribe of Australia 
a man must practice the most careful avoidance of all women who 
are included in the category of 'father's sister', of whom there are 
very many and of whom his wife's mother is one. But in other 
instances all that the application of a term implies is a certain 
general attitude rather than any more specific relation. Within 
such a category there may be a specific jural or personal relation 
to a particular individual. In many classificatory systems the terms 
for grandfather and grandmother are used in this way, as implying 
a general attitude of friendliness, relatively free from restraint, 
towards all persons to whom they are applied. Grandparents and 
grandchildren are persons with whom one can be on free and 
easy terms. This is connected with an extremely widespread, 
indeed almost universal, way of organising the relation of alternate 
generations to one another. 

In the Fox and Hopi systems all the members of the lineage 
of a grandparent are included in one category with the grand- 
parents and the attitude that is appropriate towards a grandparent 
is extended to them. This does not imply any definite set of rights 
and duties, but only a certain general type of behaviour, of a kind 
that is regarded as appropriate towards relatives of the second 
ascending generation in a great many societies not belonging to the 
Choctaw and Omaha type. 

I should have liked to discuss this further and to have dealt 
with those varieties of the Omaha type (such as the VaNdau) 
in which the mother's brother and the mother's brother's son are 
called 'grandfather'. But I have only time to draw your attention 
to a special variety of the Choctaw type which is of great interest 
in this connection. The Cherokee were divided into seven 
matrilineal clans. In the father's clan a man called all the men and 
women of his father's and all succeeding generations 'father' and 
'father's sister', and this clan and all its individual members 


had to be treated with great respect. A man could not marry a 
woman of his father's clan, and of course he could not marry into 
his own clan. In the clan of his father's father and that of his 
mother's father a man calls all the women of all generations 
'grandmother'. He thus treats, not the lineage, but the w'hole 
clan as a unity, although a clan must have numbered many 
hundreds of persons. With any woman whom he calls 'grand- 
mother' a man is allowed to be on free and easy terms. It was 
regarded as particullary appropriate that a man should marry a 
'grandmother', i.e. a woman of his mother's father's or father's 
father's clan. 

Let us now return to a brief consideration of the special 
customs of marriage that have been proposed as causes of the 
Choctaw and Omaha terminologies respectively. Marriage w4th 
the wife's brother's daughter is theoretically possible and does 
perhaps actually, though only occasionally, occur in some of the 
tribes having a system of Omaha type. Though there has been no 
marriage of this kind in the Fox tribe in recent times it is spoken 
of as a custom that formerly existed. We have seen that the marriage 
custom and the terminology fit consistently. The reason for this 
should now be easy to understand, for a little consideration will 
show that this particular marriage is an application of the prin- 
ciple of the unity of the lineage combined with the custom of the 
sororate or sororal polygyny. In the usual form of these customs 
we are concerned only with the principle of the unity of the 
sibling group. A man marries one woman of a particular sibling 
group and thereby establishes a particular relation to that group 
as a unity. The men are now permanently his brothers-in-law. 
Towards one of the women he stands in a marital relationship, and 
therefore towards the others he is conceived as standing in a 
similar relationship which may be called a quasi-marital relation- 
ship. For instance, they will regard his children as being their 
'children'. Thus it is appropriate that when he takes a second wife, 
whether before or after the death of his first, he should marry 
his wife's sister. 

I am quite aware that sororal polygyny can be attributed 
to the fact that co-wives who are sisters are less likely to quarrel 
seriously than two who are not so related, and that the sororate 
may similarly be justified by the fact that a stepmother is more 
likely to have proper affection for her stepchildren if they are the 


children of her own sister. These propositions do not conflict 
with my explanation but support it, for the principle of the 
unity of the sibling group as a structural principle is based on the 
solidarity of brothers and sisters within one family. 

When we turn to systems of the Omaha type, we see that in 
place of the unity of the sibling group we now have a unity of the 
larger group, the lineage group of three generations. When a man 
marries one woman of this group he enters into a relation with the 
group as a unity, so that all the men are now his brothers-in-law% 
and he at the same time enters into what I have called a quasi- 
marital relationship with all the women, including not only his 
wife's sisters but also his wife's brother's daughters, and in some 
systems his wife's father's sisters. The group within which, by the 
principle of the sororate, he may take a second wife without enter- 
ing into any new social bonds is thus extended to include his wife's 
brother's daughter; and the custom of marriage with this relative 
is simply the result of the application of the principle of the unity 
of the lineage in a system of patrilineal lineages. The special form 
of marriage and the special system of terminology, where they 
occur together, are directly connected by the fact that they are 
both applications of the one structural principle. There is no 
ground whatever for supposing that one is the historical cause of 
the other. 

The matter is much more complex when we come to the custom 
of marriage with the mother's brother's widow. This form of 
marriage is found associated with terminology of the Choctaw 
type in the Banks Islands, in the tribes of North-West America 
and in the Twi-speaking Akim Abuakwa. But it is also found in 
many other places w here that type of terminology does not exist. 
Nor is it correlated with matrilineal descent, for it is to be found 
in African societies that are markedly patrilineal in their institutions. 
There does not seem to be any theoretical explanation that will 
apply to all the known instances of this custom. There is no time 
on this occasion to discuss this subject by an analysis of instances. 

I must briefly refer to another theory, which goes back to 
Durkheim's review (1898) of Kohler, and by which the Choctaw 
and Omaha terminologies are explained as being the direct 
result of emphasis on matrilineal and patrilineal descent respec- 
tively. We have, fortunately, a crucial instance to which we can 
refer in this connection, in the system of the Manus of the 


Admiralty Islands, of which we have an excellent analysis by Dr. 
Margaret Mead (1934). The most important feature of the Manus 
system is the existence of patrilineal clans (called by Dr. Mead 
'gentes') and the major emphasis is on patrilineal descent. The 
solidarity of the patrilineal lineage is exhibited in many features 
of the system, but not in the terminology. However this emphasis 
on patrilineal descent is to a certain extent counterbalanced by the 
recognition of matrilineal lineages, and this does appear in the 
terminology in features that make it similar to the Choctaw type. 
Thus a single term, pinpapu, is applied to the father's father's 
sister and to all her female descendants in the female line, and a 
single term, patieyCy is applied to the father's sister and all her 
descendants in the female line. The unity of the matrilineal 
lineage is exhibited not only in the use of these terms, but also 
in the general social relation in which a person stands to the 
members of it, and is an important feature of the total complex 
kinship structure. 

One of the strange ideas that has been, and I fear still is, 
current is that if a society recognises lineage at all it can only 
recognise either patrilineal or matrilineal lineage. I believe the 
origin of this absurd notion, and its persistence in the face of 
known facts, are the result of that early hypothesis of conjectural 
history that matrilineal descent is more primitive, i.e. historically 
earlier, than patrilineal descent. From the beginning of this century 
we have been acquainted with societies, such as the Herero, in 
which both matrilineal and patrilineal lineages are recognised; 
but these were dismissed as being 'transitional' forms. This is 
another example of the way in which attachment to the method 
and hypotheses of conjectural history prevents us from seeing 
things as they are. It was this, I think, that was responsible for 
Rivers' failing to discover that the Toda system recognises 
matrilineal lineage as well as patrilineal, and that the islands of 
the New Hebrides have a system of patriUneal groups in addition 
to their matrilineal moieties. Apart from the presuppositions of the 
method of conjectural history, there is no reason why a society 
should not build its kinship system on the basis of both patriUneal 
and matrilineal lineage, and we know that there are many societies 
that do exactly this. 

In my criticism of the method of conjectural history I have 
insisted on the need for demonstration in anthropology. How then 


am I to demonstrate that my interpretation of the Choctaw- 
Omaha terminologies is the vaUd one? There are a number of 
possible arguments, but I have time for only one, which I hope may 
be considered sufficient. This is drawn from the existence of 
terminologies in which the unity of lineage or clan is exhibited, 
but which do not belong to either the Choctaw or the Omaha type; 
and I will mention one example, that of the Yaralde tribe of 
South Australia. 

The Yaralde are divided into local patrilineal totemic clans. 
A man belongs to his father's clan, and we will consider his 
relation to three other clans: those of his mother, his father's 
mother and his mother's mother. The Yaralde, like many other 
Australian tribes, such as the Aranda, have four terms for grand- 
parents, each of which is applied to both men and women. 
The term maty a is applied to the father's father and his brothers 
and sisters and to all members of a man's own clan of the second 
ascending generation. A second term, Tjaitja, is applied to the 
mother's father and his brothers and sisters, i.e. to persons of the 
mother's clan of the appropriate generation. The third term, 
mutsa, is applied not only to the father's mother and her brothers 
and sisters, but to all persons belonging to the same clan, of all 
generations and of both sexes. The clan is spoken of collectively 
as a man's mut^aurui. Similarly the term haka is applied to the 
mother's mother and her brothers and sisters and to all members 
of her clan of all generations, the clan being spoken of as a 
man's hakaurui. The structural principle here is that for the out- 
side related person the clan constitutes a unity within which 
distinctions of generation are obliterated. Compare this with the 
treatment of lineages or clans of grandparents in the Fox, Hopi 
and Cherokee systems. 

The Yaralde terminology for relatives in the mother's clan is 
shown in Fig. 13. It will be noted that the mother's brother's son 
and daughter are not called mother's brother {wano) and mother 
{fierjko) as in Omaha systems. But the son's son and daughter of 
the mother's brother are called 'mother's brother' and 'mother'. 
If we wish to explain this by a special form of marriage it would 
have to be marriage with the wife's brother's son's daughter. 
I am not certain that such a marriage would be prohibited by the 
Yaralde system, but I am quite sure that it is not a custom so 
regular as to be regarded as an effective cause in producing the 


Yaralde terininology, and it would afford no explanation whatever 
for the terminological unification of the clans of the father's 
mother and the mother's mother. The structural principle in- 
volved is obviously that of the merging of alternate generations, 
which is of such great importance in Australia, and which we have 
also seen in the Hopi system. A system very similar to the Yaralde 



mother's father 

mother's father's sister 

mother's brother 




mother's brother's son 



mother's brother's daughter 

mother's brother's 
son's son 




ne7)ko f-i) IS ng 



Fig. 13 — Yaralde 
Mother's hneage 

is found in the Ungarinyin tribe of North- West Australia, but I 
will not do more than refer to it. 

Earlier in this address I said that I would try to show you that 
the Omaha type of terminology is just as reasonable and fitting 
in those social systems in which it is found as our own terminology 
is in our system, I hope I have succeeded in doing this. On the 
basis of the elementary family and the genealogical relationships 
resulting therefrom, we English have constructed for ourselves a 
certain kinship system which meets the necessities of an ordered 
social life and is fairly self-consistent. The Fox or the Hopi have 
on the same basis constructed a relatively self-consistent system 


of a different type which provides for the needs of social cohesion 
in a different way and over a wider range. We understand the 
terminology in each instance as soon as we see it as part of an 
ordered system. The obvious connection of the Omaha termino- 
logy with the custom of marriage with the wife's brother's 
daughter is seen as a relation between two parts of a self- consistent 
working system, not as a relation of cause and effect. 

If you ask the question, 'How^ is it that the Omaha (or any 
other of the tribes we have considered) have the system that they 
do?' then it is obvious that the method of structural analysis does 
not afford an answer. But neither does conjectural history. The 
proffered but purely hypothetical explanation of the Omaha 
terminology is that it resulted from the adoption of a certain 
unusual custom of marriage. This obviously gives us no explanation 
until we know why the Omaha and other tribes came to adopt this 
custom. The only possible wa}^ of answering the question why a 
particular society has the social system that it does have is by a 
detailed study of its history over a sufficient period, generally 
several centuries. For the tribes with which we are here concerned 
the materials for such a history are entirely lacking. This is, of 
course, very regrettable, but there is nothing that v,'e can do about 
it. If you want to know how England comes to have its present 
system of constitutional monarchy and parliamentary government, 
you will go to the history books, which will give you the details of 
the growth of the system. If there were no records at all of this 
historical development, would the anthropologists think it worth 
while to spend their time in making conjectures as to what it 
might have been? 

Even when there are historical records, they only enable us 
to discover how a particular system has grown out of a somewhat 
different particular system. Thus it would be possible to write a 
historical account of the changes of the kinship system of England 
during the past ten centuries. This would take us back to the 
Teutonic bilateral sib system, as exhibited in the institution of 
wergild. But we still should not know why the Teutonic peoples 
had this kind of system, while the Romans had a different system 
of agnatic lineages. The great value of history for a science of 
society is that it gives us materials for the study of how social 
systems change. In this respect conjectural history is absolutely 


But if you ask, not how the EngHsh kinship , system or the 
EngUsh political system came into existence, but how it works at 
the present time, that is a question that can be answered by 
research of the same kind as anthropological field-work, and his- 
torical considerations are relatively, if not absolutely, unimportant. 
Such knowledge of how social systems work is of great value 
for any understanding of human life. It often has been and still 
is neglected by anthropologists who consider it their principal 
task to write the history of peoples or institutions that have no 

If you accept the analysis that I have given, but still wish to 
apply the method of conjectural history, what you have to con- 
jecture is why all the tribes that have been enumerated elected 
to construct their kinship systems on the basis of the unity of 
the lineage. 

What kind of results can we expect to obtain from the method 
of sociological analysis? Nothing, of course, that will be accept- 
able as significant by those who demand that any explanation 
of a social phenomenon must be a historical explanation, or by 
those who demand what is called psychological explanation, i.e. 
explanation in terms of the individual and his motives. I suggest 
that the results that we may reasonably expect are as follows: 

1. It will enable us to make a systematic classification of 
kinship systems. Systematic classification is an essential 
in any scientific treatment of any class of phenomena, 
and such classification must be in terms of general properties. 

2. It enables us to understand particular features of par- 
ticular systems. It does this in tw^o ways: (a) by revealing 
the particular feature as a part of an organised whole; 
{b) by showing that it is a special example of a recognisable 
class of phenomena. Thus I have tried to show that the 
Choctaw and Omaha terminologies belong to a class w'hich 
also includes the Yaralde terminology, and that these are 
all special applications of the general principle of the 
solidarity and continuity of the lineage, which appears in 
many other forms in a great number of different societies. 

3. It is the only method by which we can hope ultimately 
to arrive at valid generalisations about the nature of human 
society, i.e. about the universal characteristics of all 
societies, past, present, and future. It is, of course, such 


generalisations that are meant when we speak of sociological 

In the method of conjectural history single problems are 
usually considered in isolation. On the other hand, the method 
of structural analysis aims at a general theory, and a great many 
different facts and problems are, therefore, considered together 
and in relation to one another. It is obvious that in this address, 
inordinately long as it has been, I have only been able to touch on 
a few points in the general theory of kinship structure. I have 
dealt briefly with one or two other points in earlier publications. 
That particular part of the general theory which has occupied 
us today may be said to be the theory of the establishment of 
type relationships. I have mentioned the tendency present in 
many societies to set up a type relationship betw^een a person and 
all his relatives of the parents' generation, and the even more 
marked tendency to establish a type relationship, usually one of 
free and easy behaviour, towards the relatives of the grandparents' 
generation. I have not tried to deal with this except incidentally. 
The major part of the exposition has been concerned with tv^^o 
structural principles which are themselves examples of a more 
general structural principle or class of principles.! By the principle 
of the unity of the sibling group a type relationship is set up 
between a given person and all the members of a sibling group to 
which he is related in a certain way. It is by reference to this 
principle, I hold, that we must interpret the classificatojy ter- 
minology and such customs as the sororate and levirate.|By the 
principle of the unity of the lineage group a type relationship is set 
up between a given person and all the members of a lineage group 
to which he is related in a certain way. It is by reference to this 
principle, I hold, that we must interpret the terminologies of the 
Fox, the Hopi and the Yaralde, and other similar systems in many 
scattered parts of the world. 

If you will take the time to study two or three hundred 
kinship systems from all parts of the world you will be impressed, 
I think, by the great diversity that they exhibit. But you will also 
be impressed by the way in which some particular feature, such 
as an Omaha type of terminology, reappears in scattered and widely 
spread regions. To reduce this diversity to some sort of order 
is the task of analysis, and by its means we can, I believe, find, 
beneath the diversities, a limited number of general principles 


applied and combined in various ways. Lineage solidarity in one 
form or another is found in a majority of kinship systems. There 
is nothing at all surprising in the fact that terminologies of the 
Choctaw and Omaha type, in which it finds what may be called 
an extreme development, should be encountered in separated 
regions of America, Africa, Asia and Oceania, in many different 
families of languages, and in association with many different 
types of 'culture'. 

Last year I explained in general terms how I conceive the 
study of social structure (Radcliffe-Brown, 19406). In this address, 
by means of a particular example, I have tried to show you some- 
thing of the nature of a certain method of investigation. But do 
not think that this method can be applied only to the study of 
kinship. It is applicable in one way or another to all social 
phenomena, for it is simply the method of abstractive generalisa- 
tion by the comparison of instances, which is the characteristic 
method of the inductive sciences. 

'Why all this fuss about method?' some of you may perhaps 
ask. We cannot reach agreement as to the validity or the value of 
results unless we first reach some agreement as to objectives and 
the proper methods of attaining them. In the other natural 
sciences there is such agreement; in social anthropology there is not. 
Where we disagree, it should be the first purpose of discussion to 
define as precisely as possible the ground of difference. I have 
put my case before you, without, I hope, any unfairness towards 
those with whom I disagree. It is for you to judge which of the two 
methods that I have compared is most likely to provide that kind 
of scientific understanding of the nature of human society which 
it is the accepted task of the social anthropologist to provide for 
the guidance of mankind. 


Durkheim, E. (1898). 'Zur Urgeschichte der Ehe, Prof. J. Kohler', 

Analyses III, La Famille, Annee Sociologique, Vol. I, pp. 306-319. 
Eggan, F. (1933). 'The Kinship System and Social Organisation of the 

Western Pueblos with Special Reference to the Hopi', Ph.D. thesis, 

University of Chicago. 
GiflFord, E. W. (1916). 'Miwok Moieties', Arch, mid Ethn. Piibl., Univ. 

California, Vol. XII, No. 4. 
Gilbert, William H., Jr. (1937). 'Eastern Cherokee Social Organisation', 

in Social Anthropology of North American Tribes (ed. Fred Eggan). 

Chicago University Press, pp. 283-338. 


Kohler, J. (1897). 'Zur Urgeschichte der Ehe', Zeitschrift filr Vergleichende 

Rechtswissenschaft (Stuttgart), Bd. 11. 
Kroeber, A. L. (1909). 'Classificatory Systems of Relationship', 

J. R. Anthrop. Inst., Vol. XXXIX, pp. 77-84. 

— (1917). 'California Kinship Systems', Arch, and Ethn. Puhl. Univ. 
California, Vol. XII, No. 9. 

Mead, Margaret (1934). 'Kinship in the Admiralty Islands', Anthrop. 

Papas Amer. Mus. Nat. History, Vol. XXXIV, Pt. II, pp. 181-358. 
M'Lennan, John F. (1865). Primitive Marriage. Edinburgh: Adam & 

Charles Black. 
Morgan, Lewis H. (1871). 'The Systems of Consanguinity and Affinity', 

Smithsofiian Institution Contributions to Knozvledge, Vol. XVII. 

— (1877). Ancient Society or Researches in the Lines of Human Progress 
from Savagery to Civilisation. London: Macmillan; New York: 
Henry Holt. 

Opler, M. E. (19370). 'Chiricahua Apache Social Organisation', in Social 
Anthropology of North American Tribes (ed. Fred Eggan), Chicago 
University Press. 

— (19376). 'Apache Data Concerning the Relation of Kinship Ter- 
minology to Social Classification', Amer. Anthrop., Vol. XXXIX, 
pp. 201-212. 

Radcliffe-Brown, A. R. (1918). 'Notes on the Social Organisation of 
Australian Tribes', Pt. I, J. R. Anthrop. Inst., Vol. XLVIII, 
pp. 222-253. 

— (1924). 'The Mother's Brother in South Africa', South African 
y. Science, Vol. XXI. 

— (1930-31). 'The Social Organisation of Australian Tribes', Pts. I-III, 
Oceania, Vol. I, pp. 34-63, 206-246, 322-341, 426-456. 

— (1935). 'Patrilineal and Matrilineal Succession', Iowa Law Review, 
Vol. XX, No. 2. 

— (1940a). 'On Joking Relationships', Africa, Vol. XIII, No. 3, 
pp. 195-210. 

— (19406). 'On Social Structure', J. R. Anthrop. Inst., Vol. LXX, 
pp. 1-12. 

Rivers, W. H. R. (1907). 'On the Origin of the Classificatory System of 
Relationship', in Anthropological Essays Presented to Edward Burnett 
Tylor. Oxford: Clarendon Press. (Reprinted in Social Organisation. 
London: Kegan Paul, 1924, App. i, pp. 175-192.) 

— (1914a). History of Melanesian Society. Cambridge University Press. 

— (19146). Kinship and Social Organisation. London: London School of 

Seligman, Brenda Z. (1917). 'The Relationship Systems of the Nandi 

Masai and Thonga', Man, Vol. XVII, 46. 
Starcke, C. N. (1889). The Primitive Family (The International Scientific 

Series, Vol. LXVI). London: Kegan Paul. 
Stewart, Dugald (1795). Introduction to Essays of Adam Smith. 
Tax, Sol (1937). 'The Social Organisation of the Fox Indians', in Social 

Anthropology of North American Tribes (ed. Fred Eggan). Chicago 

University Press, pp. 241-282. 



THE publication of Mr. F. J. Pedler's note^ on what are 
called 'joking relationships', following on two other papers 
on the same subject by Professor Henri Labouret^ 
and Mademoiselle Denise Paulme,* suggests that some general 
theoretical discussion of the nature of these relationships may be 
of interest to readers of Africa.^ 

What is meant by the term 'joking relationship' is a relation 
between two persons in which one is by custom permitted, and in 
some instances required, to tease or make fun of the other, who in 
turn is required to take no offence. It is important to distinguish 
[ two main varieties. In one the relation is symmetrical; each of the 
two persons teases or makes fun of the other. In the other variety 
the relation is asymmetrical; A jokes at the expense of B and B 
accepts the teasing good humouredly but without retaliating; or 
A teases B as much as he pleases and B in return teases A only a 
little. There are many varieties in the form of this relationship in 
different societies. In some instances the joking or teasing is only 
verbal, in others it includes horse-play; in some the joking 
includes elements of obscenity, in others not. 

Standardised social relationships of this kind are extremely 

C widespread, not only in Africa but also in Asia, Oceania and 
North America. To arrive at a scientific understanding of the 
phenomenon it is necessary to make a wide comparative study. 
Some material for this now exists in anthropological literature, 
though by no means all that could be desired, since it is un- 
fortunately still only rarely that such relationships are observed 
and described as exactly as they might be. 

^ Reprinted from Africa, Vol. XIII, No. 3, 1940, pp. 195-210. 

^ 'Joking Relationships in East Africa', Africa, Vol. XIII, p. 170. 

^ 'La Parent^ a Plaisanteries en Afrique Occidentale', Africa, Vol. II, p. 244. 

* 'Parent^ a Plaisanteries et Alliance par le Sang en Afrique Occidentale', 
Africa, Vol. XII, p. 433. 

^ Professor Marcel Mauss has published a brief theoretical discussion of the 
subject in the Annuaire de VEcole Pratique des Hautes Etudes, Section des 
Sciences religietises , 1927-8. It is also dealt with by Dr. F. Eggan in Social 
Anthropology of North American Tribes, 1937, pp. 75-81. 



The joking relationship is a pecuHar combination of friendliness 
and antagonism- The behaviour is^ such that in any other social 
context it would express and arouse hostility; but it is not meant 
seriously and must not be taken seriously. There is a pretence of 
hostility and a real friendliness. To put it in another way, the 
relationship is one of j3ermitted_^isrespect. Thus any complete 
theory of it must be part of, or consistent with, a theory of the place 
of resp^ect in^spcial relations and in social life generally. But this is^.. 
a very wide and very important sociological problem; for it is / 
evident that the whole maintenance of a social order depends upon / 
the appropriate kind and degree of respect being shown towards I 
certain persons, things and ideas or symbols. 

Examples of joking relationships between relatives by marriage . 
are very commonly found in Africa and in other parts of the world. 
Thus Mademoiselle Paulme^ records that among the Dogon a 
man stands in a joking relationship to his wife's sisters and their 
daughters. Frequently the relationship holds betw'een a man and 
both the brothers and sisters of his wife. But in some instances 
there is a distinction whereby a man is on joking terms with his 
wife's younger brothers and sisters but not with those who are 
older than she is. This joking with the wife's brothers and sisters 
is usually associated with a custom requiring extreme respect, 
often partial or complete avoidance, between a son-in-law and his 
wife's parents. 2 

The kind of structural situation in which the associated 
customs of joking and avoidance are found may be described as 
follows. A marriage involves a readjustment of the social struc- _-n 
.ture whereby the woman's relations with her family are greatly < 
modified and she enters into a new and very close relation with her 
husband. The latter is at the same time brought into a special 
relation with his wife's family, to which, however, he is an outsider. 
For the sake of brevity, though at the risk of over-simplification, 
we will consider only the husband's relation to his wife's family. 
The relation can be c'escribed as involving both attachment and 
separation, both social conjunction and social disjunction, if I 

1 Africa, Vol. XII, p. 438. 

^ Those who are not familiar with these widespread customs will find descrip- 
tions in Junod, Life of a South African Tribe, NeuchStel, Vol. I, pp. 229-37, ^nd 
in Social Aytthropology of North American Tribes, edited by F. Eggan, Chicago, 
1937, PP- 55-7- 


may use the terms. The man has his own definite position in the 
social structure, determined for him by his birth into a certain 
family, Hneage or clan. The great body of his rights and duties and 
the interests and activities that he shares with others are the result 
of his position. Before the marriage his wife's family are outsiders 
for him as he is an outsider for them. This constitutes a social 
disjunction which is not destroyed by the marriage. The social 
conjunction results from the continuance, though in altered form, 
of the wife's relation to her family, their continued interest in her 
and in her children. If the wife were really bought and paid for, 
as ignorant persons say that she is in Africa, there would be ho 
place for any permanent close relation of a man with his w^ife's 
family. But though slaves can be bought, wives cannot. 

Social disjunction implies divergence of interests and therefore 
the possibility of conflict and hostility, while conjunction requires 
the avoidance of strife. How can a relation which combines the 
)<^'~~ two be given a stable, ordered form? There are two ways of doing 
this. One is to maintain between two persons so related an extreme 
mutual respect and a limitation of direct personal contact. This is 
exhibited in the very formal relations that are, in so many societies, 
characteristic of the behaviour of a son-in law on the one side and 

this wife's father and mother on the other. In its most extreme form 
there is complete avoidance of any social contact between a man 
and his mother-in-law. 

This avoidance must not be mistaken for a sign of hostility. 

One does, of course, if one is wise, avoid having too much to do 

with one's enemies, but that is quite a different matter. I once 

asked an Australian native why he had to avoid his mother-in-law, 

and his reply was, 'Because she is my best friend in the world; she 

r has given me my wife'. The mutual respect between son-in-law 

\ and parents-in-law is a mode of friendship. It prevents conflict 

\ that might arise through divergence of interest. 

; The alternative to this relation of extreme mutual respect and 

V /restraint is the joking relationship, one, that is, of mutual dis- 

~^ respect and licence. Any serious hostility is prevented by the 

playful antagonism of teasing, and this in its regular repetition is 

a constant expression or reminder of that social disjunction which 

is one of the essential components of the relation, while the social 

conjunction is maintained by the friendliness that takes no 

offence at insult. 


The discrimination within the wife's family between those 
who have to be treated with extreme respect and those with 
whom it is a duty to be disrespectful is made on the basis of 
generation and sometimes of seniority within the generation. 
The usual respected relatives are those of the first ascending 
generation, the wife's mother and her sisters, the wife's father 
and his brothers, sometimes the wife's mother's brother. The 
joking relatives are those of a person's own generation; but very 
frequently a distinction of seniority within the generation is made; 
a wife's older sister or brother may be respected while those 
younger will be teased. 

In certain societies a man may be said to have relatives by 
marriage long before he marries and indeed as soon as he is born 
into the world. This is provided by the institution of the required 
or preferential marriage. We will, for the sake of brevity, consider 
only one kind of such organisations. In many societies it is re- 
garded as preferable that a man should marry the daughter of 
his mother's brother; this is a form of the custom known as cross- 
cousin marriage. Thus his female cousins of this kind, or all those 
women whom by the classificatory system he classifies as such, 
are potential wives for him, and their brothers are his potential 
brothers-in-law. Among the Ojibwa Indians of North America, 
the Chiga of Uganda, and in Fiji and New Caledonia, as well as 
elsewhere, this form of marriage is found and is accompanied by 
a joking relationship between a man and the sons and daughters 
of his mother's brother. To quote one instance of these, the 
following is recorded for the Ojibwa. 'When cross-cousins meet 
they must try to embarrass one another. They "joke" one another, 
making the most vulgar allegations, by their standards as well 
as ours. But being "kind" relations, no one can take oflFence. 
Cross-cousins who do not joke in this way are considered boorish, 
as not playing the social game.'^ 

The joking relationship here is of fundamentally the same kind 
as that already discussed. It is established before marriage and is 
continued, after marriage, with the brothers- and sisters-in-law. 

In some parts of Africa there are joking relationships that have 
nothing to do with marriage. Mr. Pedler's note, mentioned above, 
refers to a joking relationship between two distinct tribes, the 

^ Ruth Landes in Mead, Co-operation and Competition among Primitive 
Peoples, 1937, p. 103. 


Sukuma and the Zaramu, and in the evidence it was stated that 
there was a similar relation between the Sukuma and the Zigua 
and between the Ngoni and the Bemba. The woman's evidence 
suggests that this custom of rough teasing exists in the Sukuma 
tribe between persons related by marriage, as it does in so many 
other African tribes.^ 

While a joking relationship between two tribes is apparently 
rare, and certainly deserves, as Mr. Pedler suggests, to be carefully 
investigated, a similar relationship between clans has been observed 
in other parts of Africa. It is described by Professor Labouret and 
Mademoiselle Paulme in the articles previously mentioned, and 
amongst the Tallensi it has been studied by Dr. Fortes, who will 
deal with it in a forthcoming publication. ^ 

The two clans are not, in these instances, specially connected 
by intermarriage. The relation between them is an alliance 
involving real friendliness and mutual aid combined with an 
appearance of hostility. 

The general structural situation in these instances seems to be 
as follows. The individual is a member of a certain defined group, 
a clan, for example, within which his relations to others are 
defined by a complex set of rights and duties, referring to all the 
major aspects of social life, and supported by definite sanctions. 
There may be another group outside his own which is so linked 
with his as to be the field of extension of jural and moral relations 
of the same general kind. Thus, in East Africa, as we learn from 
Mr. Pedler's note, the Zigua and the Zaramu do not joke with 
one another because a yet closer bond exists between them since 
they are ndugu (brothers). But beyond the field within which social 
relations are thus defined there lie other groups with which, since 

^ Incidentally it may be said that it was hardly satisfactory for the magistrate 
to establish a precedent whereby the man, who was observing what was a 
permitted and may even have been an obligatory custom, was declared guilty of 
common assault, even with extenuating circumstances. It seems quite possible 
that the man may have committed a breach of etiquette in teasing the woman 
in the presence of her mother's brother, for in many parts of the world it is 
regarded as improper for two persons in a joking relationship to tease one another 
(particularly if any obscenity is involved) in the presence of certain relatives 
of either of them. But the breach of etiquette would still not make it an assault. 
A little knowledge of anthropology would have enabled the magistrate, by 
putting the appropriate questions to the witnesses, to have obtained a fuller 
understanding of the case and all that was involved in it. 

^ Fortes, M., The Dynamics of Clanship among the Tallensi. Oxford Univer- 
sity Press, 1945. 


they are outsiders to the individual's own group, the reiationj 
involve^possible or actual hostility.. In any fixed relations between > 
the members of two such groups the separateness of the groups 
must be recognised. It is precisely this separateness which is not 
merely recognised but emphasised when a joking relationship isi 
established. The show of hostility, the perpetual disrespect, is a 
continual expression of that social disjunction which is an essential ) 
part of the whole structural situation, but over which, without 
destroying or even weakening it, there is provided the social/" / 
conjunction of friendliness and mutual aid. ' .^ 

The theory that is here put forward, therefore, is that both the '^ 
joking relationship which constitutes an alliance between clans or ) ^ 
tribes, and that between relatives by marriage, are modes of j _ , 
organising a definite and stable system of social behaviour in 
which conjunctive and disjunctive components, as I have called ; 
them, are maintained and combined. 

To provide the full evidence for this theory by following out 
its implications and examining in detail its application to different 
instances would take a book rather than a short article. But some 
confirmation can perhaps be offered by a consideration of the 
way in which respect and disrespect appear in various kinship 
relations, even though nothing more can be attempted than a 
very brief indication of a few significant points. "^/' 

In studying a kinship system it is possible to distinguish the j 
different relatives by reference to the kind and degree of respect 
that is paid to them.^ Although kinship systems vary very much -- ' 
in their details there are certain principles which are found to be 
very widespread. One of them is that by which a person is re- 
quired to show a marked respect to relatives belonging to the 
generation immediately preceding his ow^n. In a majority of 
societies the father is a relative to whom marked respect must be 
shown. This is so even in many so-called matrilineal societies, 
i.e. those which are organised into matrilineal clans or lineages. 
One can very frequently observe a tendency to extend this 
attitude of respect to all relatives of the first ascending generation 
and, further, to persons who are not relatives. Thus in those 
Er .!~ <: p{rc. r /c o i- t -•: ''■> ^ "s^?;- ' 

^ See, for example, the kinship systems described in Social Anthropology of 
North American Tribes, edited by Fred Eggan, University of Chicago Press, 
1937; and Margaret Mead, 'Kinship in the Admiralty Islands', Anthropological 
Papers of the American Museum of Natural History, Vol. XXXIV, pp. 243-56. 


tribes of East Africa that are organised into age-sets a man is 
required to show special respect to all men of his father's age- 
set and to their wives. 

The social function of this is obvious. The social tradition is 
handed down from one generation to the next. For the tradition 
to be maintained it must have authority behind it. The authority is 
therefore normally recognised as possessed by members of the 
preceding generation and it is they who exercise discipline. As a 
result of this the relation between persons of the two generations 
usually contains an element of inequality^ the parents and those 
of their generation being in a position of superiority over the 
children who are subordinate to them. The unequal relation 
between a father and his son is maintained by requiring the latter 
to show respect to the former. The relation is asymmetrical. 

When we turn to the relation of an individual to his grand- 
parents and their brothers and sisters we find that in the majority 
of human societies relatives of the second ascending generation 
are treated with very much less respect than those of the first 
ascending generation, and instead of a marked inequality there is a 
tendency to approximate to a friendly equality. 

Considerations of space forbid any full discussion of this feature 
of social structure, which is one of very great importance. There 
are many instances in which the grandparents and their grand- 
children are grouped together in the social structure in opposition 
to their children and parents. An important clue to the under- 
standing of the subject is the fact that in the flow of social life 
through time, in which men are born, become mature and die, 
the grandchildren replace their_grandparents. S© s/m/^^aj- 

In many societies there is an actual joking relationship, usually 
of a relatively mild kind, between relatives of alternate genera- 
tions. Grandchildren make fun of their grandparents and of those 
who are called grandfather and grandmother by the classificator}' 
system of terminology, and these reply in kind. 

Grandparents and grandchildren are united by kinship; they are 
separated by age and by the social difference that results from the 
fact that as the grandchildren are in process of entering into full 
participation in the social life of the community the grandparents 
are gradually retiring from it. Important duties towards his re- 
latives in his own and even more in his parents' generation impose 
upon an individual many restraints; but with those of the second 


ascending generation, his grandparents and collateral relatives, 
there can be, and usually is, established a relationship of simple 
friendliness relatively free from restraint. In this instance also, 
it is suggested, the joking relationship is a method of ordering a 
relation which combines social conjunction and disjunction. 

This thesis could, I believe, be strongly supported if not 
demonstrated by considering the details of these relationships. 
There is space for only one illustrative point. A very common 
form of joke in this connection is for the grandchild to pretend 
that he wishes to marry the grandfather's wife, or that he intends 
to do so when his grandfather dies, or to treat her as already being 
his wife. Alternatively the grandfather may pretend that the wife 
of his grandchild is, or might be, his wife,^ The point of the joke 
is the pretence at ignoring the difference of age between the 
grandparent and the grandchild. 

In various parts of the world there are societies in which a 
sister's son teases and otherwise behaves disrespectfully towards 
his mother's brother. In these instances the joking relationship 
seems generally to be asymmetrical. For example the nephew- 
may take his uncle's property but not vice versa; or, as amongst 
the Nama Hottentots, the nephew may take a fine beast from 
his uncle's herd and the uncle in return takes a wretched beast 
from that of the nephew. ^ 

The kind of social structure in which this custom of privileged 
disrespect to the mother's brother occurs in its most marked forms, 
for example the Thonga of South-East Africa, Fiji and Tonga in 
the Pacific, and the Central Siouan tribes of North America, is 
characterised by emphasis on patrilineal lineage and a marked 
distinction between relatives through the father and relatives 
through the mother. 

In a former publication^ I offered an interpretation of this 
custom of privileged familiarity towards the mother's brother. 
Briefly it is as follows. For the continuance of a social system 
children require to be cared for and to be trained. Their care 
demands affectionate and unselfish devotion; their training! 

^For examples see Labouret, Les Tribus du Rameau Lobi, 1931, p. 248, and 
Sarat Chandra Roy, The Oraons of Chota Nagpur, Ranchi, 1915. PP- 352-4- 

^ A. Winifred Hoernl6, 'Social Organisation of the Nama Hottentot'; 
American Anthropologist, N.s., Vol. XXVII, 1925, pp. 1-24. 

* 'The Mother's Brother in South Africa', South African Journal of Science, 
Vol. XXI, 1924. See Chapter I. 




requires that they shall be subjected to discipline. In the societies 
with which we are concerned there is something of a division of 
function between the parents and other relatives on the two sides, 
i The control and discipline are exercised chiefly by the father and 
,' his brothers and generally also by his sisters; these are relatives 
who must be respected and obeyed. It is the mother who is 

' primarily responsible for the affectionate care; the mother and her 
brothers and sisters are therefore relatives who can be looked to 

I for assistance and indulgence. The mother's brother i^ called 

Ij-'male mother' in Tonga and in some South African tribes. 

I believe that this interpretation of the special position of the 
mother's brother in these societies has been confirmed by further 
field work since I wrote the article referred to. But I was quite 
aware at the time it was written that the discussion and interpre- 
tation needed to be supplemented so as to bring them into line with 
a general theory of the social functions of respect and disrespect. 

The joking relationship with the mother's brother seems to 
fit well with the general theory of such relationships here outlined. 
A person's most important duties and rights attach him to his 
paternal relatives, living and dead. It is to his patrilineal lineage 
or clan that he belongs. For the members of his mother's lineage 
he is an outsider, though one in whom they have a very special 
and tender interest. Thus here again there is a relation in which 
there is both attachment, or conjunction, and separation, or 
disjunction, between the two persons concerned. 

But let us remember that in this instance the relation is 

, asymmetrical.^ The nephew is disrespectful and the uncle 
accepts the disrespect. There is inequality and the nephew is the 

, superior. This is recognised by the natives themselves. Thus in 
Tonga it is said that the sister's son is a 'chief (eiki) to his mother's 
brother, and Junod^ quotes a Thonga native as saying 'The 
uterine nephew is a chief! He takes any liberty he likes with his 
maternal uncle'. Thus the joking relationship with the uncle 
does not merely annul the usual relation between the two genera- 
tions, it reverses it. But while the superiority of the father and the 

^ There are some societies in which the relation between a mother's brother 
and a sister's son is approximately symmetrical, and therefore one of equality. 
This seems to be so in the Western Islands of Torres Straits, but we have no 
information as to any teasing or joking, though it is said that each of the two 
relatives may take the property of the other. 

- Life of a South African Tribe, Vol. I, p. 255. 


father's sister is exhibited in the respect that is shown to them, the 
nephew's superiority to his mother's brother takes the opposite 
form of permitted disrespect. 

It has been mentioned that there is a widespread tendency to 
feel that a man should show respect towards, and treat as social 
' superiors, his relatives in the generation preceding his own, and 
the custom of joking with, and at the expense of, the maternal 
I uncle clearl)' conflicts with this tendency. This conflict between 
principles of behaviour helps us to understand what seems at first 
sight a very extraordinary feature of the kinship terminology of the 
Thonga tribe and the VaNdau tribe in South-East Africa. Amongst 
the Thonga, although there is a term malutne (= male mother) for 
the mother's brother, this relative is also, and perhaps more 
frequently, referred to as a grandfather (kokwana) and he refers 
to his sister's son as his grandchild {ntukulu). In the VaNdau 
tribe the mother's brother and also the mother's brother's son are 
called 'grandfather' {tetekidii, literally 'great father') and their 
wives are called 'grandmother' (mbiya), while the sister's son and 
the father's sister's son are called 'grandchild' (miizukulu). 

This apparently fantastic w'ay of classifying relatives can be 
interpreted as a sort of legal fiction whereby the male relatives 
of the mother's lineage are grouped together as all standing towards 
an individual in the same general relation. Since this relation is 
one of privileged familiarity on the one side, and solicitude and 
indulgence on the other, it is conceived as being basically the 
one appropriate for a grandchild and a grandfather. This is 
indeed in the majority of human societies the relationship in W'hich 
this pattern of behaviour most frequently occurs. By thisjegal 
Jiction the mother's brother ceases to belong to the first ascending 
generation, of which it is felt that the members ought to be 

It may be worth while to justify this interpretation by con- 
sidering another of the legal fictions of the VaNdau terminology. 
In all these south-eastern Bantu tribes both the father's sister. and 
the sister, particularly the elder sister, are persons who must be 
treated with great respect. They are also both of them members of 
a man's own patrilineal lineage. Amongst the VaNdau the father's 
sister is called 'female father' (tetadji) and so also is the sister.^ 

^ For the kinship terminology of the VaNdau see Boas, 'Das Verwandtschafts- 
system der Vandau', in Zeitschrift fiir Etlmologie, 1922, pp. 41-51. 



Thus by the fiction of terminological classification the sister is 
placed in the father's generation, the one that appropriately 
includes persons to whom one must exhibit marked respect. 

In the south-eastern Bantu tribes there is assimilation of two 
kinds of joking relatives, the grandfather and the mother's brother. 
It may help our understanding of this to consider an example 
in which the grandfather and the brother-in-law are similarly 
grouped together. The Cherokee Indians of North America, 
probably numbering at one time about 20,000, were divided into 
seven matrilineal clans. ^ A man could not marry a woman of his 
own clan or of his father's clan. Common membership of the same 
clan connects him with his brothers and his mother's brothers. 
Towards his father and all his relatives in his father's clan of his 
own or his father's generation he is required by custom to show a 
marked respect. He applies the kinship term for 'father' not only 
to his father's brothers but also to the sons of his father's sisters. 
Here is another example of the same kind of fiction as described 
above; the relatives of his own generation whom he is required 
to respect and who belong to his father's matrilineal lineage are 
spoken of as though they belonged to the generation of his parents. 
The body of his immediate kindred is included in these two 
clans, that of his mother and his father. To the other clans of the 
tribe he is in a sense an outsider. But with two of them he is 
connected, namely with the clans of his two grandfathers, his 
father's father and his mother's father. He speaks of all the 
members of these two clans, of whatever age, as 'grandfathers' 
and 'grandmothers'. He stands in a joking relationship with all 
of them. When a man marries he must respect his wife's parents 
but jokes with her brothers and sisters. 

The interesting and critical feature is that it is regarded as 
particularly appropriate that a man should marry a woman whom 
he calls 'grandmother', i.e. a member of his father's father's 
clan or his mother's father's clan. If this happens his v^ ife's brothers 
and sisters, whom he continues to tease, are amongst those whom 
he previously teased as his 'grandfathers' and 'grandmothers'. 
This is analogous to the widely spread organisation in which a 
man has a joking relationship with the children of his mother's 
brother and is expected to marry one of the daughters. 

^ For an account of the Cherokee see Gilbert, in Social Anthropology of 
North American Tribes, pp. 285-338. 


It ought perhaps to be mentioned that the Cherokee also have 
a one-sided joking relationship in which a man teases his father's 
sister' S-husband. The same custom is found in Mota of the Bank 
Islands. In both instances we have a society organised on a 
matrilineal basis in which the mother's brother is respected, the 
father's sister's son is called 'father' (so that the father's sister's 
husband is the father of a 'father'), and there is a special term 
for the father's sister's husband. Further observation of the 
societies in which this custom occurs is required before we can be 
sure of its interpretation. I do not remember that it has been 
reported from any part of Africa. 

What has been attempted in this paper is to define in the most ~) 
general and abstract terms the kind of structural situation in which / ; 
we may expect to find well-marked joking relationships. We have- 
been dealing with societies in which the basic social structure is 
provided by kinship. By reason of his birth or adoption into a 
certain position in the social structure an individual is connected 
with a large number of other persons. With some of them he 
finds himself in a definite and specific jural relation, i.e. one which (27 
can be defined in terms of rights and duties. Who these persons 
will be and what will be the rights and duties depend on the form 
taken by the social structure. As an example of such a specific 
jural relation we may take that which normally exists between 
a father and son, or an elder brother and a younger brother. 
Relations of the same general type may be extended over a con- 
siderable range to all the members of a lineage or a clan or an 
age-set. Besides these specific jural relations which are defined 
not only negatively but also positively, i.e. in terms of things that 
must be done as well as things that must not, there are general ^ 
jural relations which are expressed almost entirely in terms of 
prohibitions and which extend throughout the whole political 
society. It is forbidden to kill or wound other persons or to take or 
destroy their property. Besides these two classes of social relations\ 
there is another, including many very diverse varieties, which can I 
perhaps be called relations of alliance or consociation. For( {J^ 
example, there is a form of alliance of very great importance in ' 
many societies, in which two persons or two groups are connected 
by an exchange of gifts or services.^. Another example is provided J 

^ See Mauss, 'Essai sur le Don', Annee Sociologique, Nouvelle S6rie, tome I, 
pp. 30-186. 


by the institution of blood-brotherhood which is so widespread 
in Africa. 

The argument of this paper has been intended to show that the 
joking relationship is one special form of alliance in this sense. An~ 
alliance by exchange of goods and services may be associated with 
a joking relationship, as in the instance recorded by Professor 
Labouret.^ Or it may be combined with the custom of avoidance. 
Thus in the Andaman Islands the parents of a man and the parents 
of his wife avoid all contact with each other and do not speak; at 
the same time it is the custom that they should frequently exchange 
presents through the medium of the younger married couple. 
But the exchange of gifts may also exist without either joking 
or avoidance, as in Samoa, in the exchange of gifts between 
the family of a man and the family of the woman he marries 
or the very similar exchange between a chief and his 'talking 

So also in an alliance by blood-brotherhood there may be a 
joking relationship as amongst the Zande;^ and in the somewhat 
similar alliance formed by exchange of names there may also be 
mutual teasing. But in alliances of this kind there may be a 
relation of extreme respect and even of avoidance. Thus in the 
Yaralde and neighbouring tribes of South Australia two boys 
belonging to communities distant from one another, and therefore 
more or less hostile, are brought into an alliance by the exchange of 
their respective umbilical cords. The relationship thus established 
is a sacred one; the two boys may never speak to one another. 
But when they grow up they enter upon a regular exchange of 
gifts, which provides the machinery for a sort of commerce 
between the two groups to which they belong. 

Thus the four modes of alliance or consociation, (i) through 
intermarriage, (2) by exchange of goods or services, (3) by blood- 
brotherhood or exchanges of names or sacra, and (4) by the joking 
relationship, may exist separately or combined in several different 
ways. The comparative study of these combinations presents 
a number of interesting but complex problems. The facts recorded 
from West Africa by Professor Labouret and Mademoiselle 
Paulme afford us valuable material. But a good deal more intensive 

^ Africa, Vol. II, p. 245. 

^ Evans-Pritchard, 'Zande Blood-brotherhood', Africa, Vol. VI, 1933, 
pp. 369-401. 


field research is needed before these problems of social structure 
can be satisfactorily dealt with. 

What I have called relations by alliance need to be compared 
with true contractual relations. The latter are specific jural 
relations entered into by two persons or two groups, in which 
either party has definite positive obligations towards the other, 
and failure to carry out the obligations is subject to a legal 
sanction. In an alliance by blood-brotherhood there are general 
obligations of mutual aid, and the sanction for the carrying out 
of these, as shown by Dr. Evans- Pritchard, is of a kind that can 
be called magical or ritual. In the alliance by exchange of gifts 
failure to fulfil the obligation to make an equivalent return for a 
gift received breaks the alliance and substitutes a state of hostility 
and may also cause a loss of prestige for the defaulting party. 
Professor Mauss^ has argued that in this kind of alliance also 
there is a magical sanction, but it is very doubtful if such is 
always present, and even when it is it may often be of secondary 

The joking relationship is in some ways the exact opposite of 
a contractual relation. Instead of specific duties to be fulfilled there —1 
is privileged disrespect and freedom or even licence, and the only 
obligation is not to take offence at the disrespect so long as it is 
kept within certain bounds defined by custom, and not to go / 
beyond those bounds. Any default in the relationship is like ai 
breach of the rules of etiquette; the person concerned is regarded i 
as not knowing how to behave himself. 

In a true contractual relationship the two parties are conjoined 
by a definite common interest in reference to which each of them 
accepts specific obligations. It makes no difference that in other 
matters their interests may be divergent. In the joking relation- 
ship and in some avoidance relationships, such as that between a 
man and his wife's mother, one basic determinant is that the 
social structure separates them in such a way as to make many of 
their interests divergent, so that conflict or hostility might result. 
The alliance by extreme respect, by partial or complete avoidanceL ^ 
prevents such conflict but keeps the parties conjoined. The ^ 
alliance by joking does the same thing in a different way. ' 

All that has been, or could be, attempted in this paper is to 
show the place of the joking relationship in a general comparative 

^ *Essai sur le Don*. 


Study of social structure. What I have called, provisionally, 
relations of consociation or alliance are distinguished from the 
relations set up by common membership of a political society 
which are defined in terms of general obligations, of etiquette, or 
morals, or of law. They are distinguished also from true con- 
tractual relations, defined by some specific obligation for each 
contracting party, into which the individual enters of his own 
volition. They are further to be distinguished from the relations 
set up by common membership of a domestic group, a lineage or a 
clan, each of which has to be defined in terms of a whole set of 
socially recognised rights and duties^JRelations of consoci ation 
can only exist between individuals or groups which are in some 
way socially separated. 

This paper deals only with formalised or standardised Joking 
relations. Teasing or making fun of other persons is of course a 
common mode of behaviour in any human society. It tends to 
occur in certain kinds of social situations. Thus I have observed 
in certain classes in English-speaking countries the occurrence 
of horse-play between young men and women as a preliminaiy to 
courtship, very similar to the way in which a Cherokee Indian 
jokes with his 'grandmothers'. Certainly these unformalised modes 
of behaviour need to be studied by the sociologist. For the purpose 
of this paper it is sufficient to note that teasing is always^ a coin- 
pound of friendliness and antagonism. 

The scientific explanation of the institution in the particular 

form in which it occurs in a given society can only be reached by an 

1 intensive study which enables us to see it as^ particular example of 

\ a widespread phenomenon of a definite class, This means that the 

whole social structure has to be thoroughly examined in order that 

the particular form and incidence of joking relationships can be 

understood as part of a consistent system. If it be asked why that 

society has the structure that it does have, the only possible 

i answer w^ould lie in its history. When the histor}^ is unrecorded, 

I as it is for the native societies of Africa, we can only indulge in 

conjecture, and conjecture gives us neither scientific nor historical 


^ The general theory outlined in this paper is one that I have presented in 
lectures at various universities since 1909 as part of the general study of the forms 
of social structure. In arriving at the present formulation of it I have been 
helped by discussions with Dr. Meyer Fortes. 



PROFESSOR Griaule's article on 'L' Alliance cathartique' in 
Africa of October 1948 raises a methodological point of 
considerable importance. If we wish to understand 
custom or institution that we find in a particular society there ar^ 
two ways of dealing with it. One is to examine the part it plays (T) 
in the,system or complex of customs and institutions in which it_ 
is found and the meaning that it has within this complex for th£ 
peopIeThemselves./ Professor Griaule deals in this way with the 
custom by which the Bozo and the Dogon exchange insults with 
each other. He considers it as an element in a complex of customs, 
institutions, myths, and ideas to which the Dogon themselves 
refer by the term wow^w/^ He shows us also what meaning the 
natives themselves attribute to this exchange of insults (p. 253). 
As a piece of analysis the article is admirable, and is a most im- 
portant contribution to our growing knowledge of West African 

But there is another method open to us, namely, to make a 
wide cornparative study_of all those types of social relationship (__^ 
in which two persons are by custom permitted, or even required, 
to use speech or behaviour which in other relationships would be 
grievously offensive. To the use of this method it would seem 
that Professor Griaule objects. Referring to what has already 
been written on the comparative study of what are called 'joking 
relationships' or parentis a. plaisanterie he writes: 'Nous adoptons^'V _ 
vis-a-vis travaux parus sur cette question, une attitude negative.' J 

Ethnographers had reported from North. America, Oceania, 
and_ Africa instances of a custom by which persons standing 
in certain relationships resulting either from kinship, or more 
usually from marriage, were permitted or required to behave 
towards one another in a disrespectful or insulting way at which 
no offence might be taken. Such relationships came to be called 
'jo king rela tionships', admittedly not a very good name. The most 

^ Reprinted from Africa, Vol. XIX, 1949, pp. 133-140. 


numerous and widespread examples of this custom were in the 
relationship of a man to the brothers and sisters of his wife. 
But it was also found in some instances between cross-cousins, 
between mother's brother and sister's son, and in a somewhat 
milder form between grandparents and grandchildren. There 
thus arose a problem of comparative sociology: What is there in all 
these relationships that makes this type of behaviour app reprice, 
meaningful, and functional? 

One of the first facts that strikes the sociological enquirer is 
that the custom of 'joking' with the wife's brothers and sisters is 
very commonly associated with a custom of strict avoidance of the 
wife's mother, frequently of the wife's father, and more occasion- 
ally the wife's mother's brother. Since it is clear that the avoidance 
custom and the joking custom are direct contraries, or polar^ 
opposites, the problem immediately became one of dealing with 
both these types of custom. And this in turn made it necessary 
to consider certain other kinds of relationships. 

I became interested in this whole set of problems in 1908 when 
I was trying to find an explanation of customs of avoidance in the 
Andaman Islands. There the parents of a man and the parents 
of his wife must avoid each other. Their relationship is described 
by the term aka-yat, from a stem meaning 'forbidden' and a prefix 
referring to the mouth, and, therefore, to speech. Persons in such 
a relationship might not speak to each other. On the other hand 
I was told that they will regularly send each other presents. The 
explanation given by the Andamanese is: 'They are great friends 
because their children have married.' This conception of avoid- 
ance relationships as relationships of friendship I have found 
elsewhere. Thus in Australia, where a man carefully avoids all 
social contact with his wife's mother, I have more than once 
been told that she is the greatest friend he has since she has pro- 
vided him with a wife. Again, the joking relationship is commonly 
referred to as one of friendship. 'I can tease my mother's brother 
and take his property because we are great friends; I am the son 
of his sister.' 'I can joke with my grandfather or grandmother, and 
they will joke with me because we are great friends.' 

What does 'friendship' mean in these contexts? It is clearly 
something different from the relationship of solidarity and mutual 
help between two brothers or a father and son. On the basis of 
comparative analysis it seems to me that the assertion of 'friend- 


ship' means an obligation for the two persons not to enter into 1 
open quarrel or conflict with each other. It is sufficiently evident 1 
that one way of obviating open conflict between two persons is }. 
for them to avoid one another or treat each other with very marked,^ 
respect. I think it is also fairly evident that a relationship in which! ■ 
insults are exchanged and there is an obligation not to take them j 
seriously, is one which, by means of sham conflicts, avoids real/ 
ones. ,.-^ 

This theory can be supported by reference to customs of 
other kinds, of which, to economise space, I ^^ ill only mention two 
that are typical of one kind. In the Andamans I was told that 
two men who were initiated together at the same ceremony of 
initiation would be forbidden thereafter to speak to one another, 
but would regularly exchange gifts. Again the explanation was: 
'They are great friends.' In South Australia there was a custom 
whereby two boys, born about the same time in two clans that 
were normally hostile, were united into a special relationship by the 
exchange of that portion of the umbilical cord which remains 
on the infant and later falls off. The two men who stand in this 
relationship may never speak to each other, but each may visit in 
safety the clan of the other carrying gifts to his friend and receiving 
gifts in exchange. Again the relationship is described as one of y 
great friendship; through it each of the persons is safe in what / 
would otherwise be hostile territory. -^ 

A careful examination of a great many instances from all over? 
the world seems to me to justify the formulation of a genexali- 
_theory. But these special forms of 'friendship' can, of course/ 
only be fully dealt with in terms of a study of forms of social 
relationship in general, and this is not the place in which to take 
up that very wide subject. Some social relationships are required 
by custom to be based on respect, of difl^crent degrees and ex- 
pressed in diff^erent ways; others are such as to permit a certain 
degree of familiarity, and in extreme cases of licence. The_rules 
of etiquette are one method of standardising these features of 
social relations. The respect required of a son to his father in many 
African tribes must be exhibited in this way. The avoidance 
relationship is in one sense an extreme form of respect, while the 
joking relationship is a form of familiarity, permitting disrespectful 
behaviour, and in extreme instances, licence. It is, for example, a 
relationship in which, in some cases, obscenity may be freely 


indulged in, as between the Dogon and the Bozo. Obscene talk, in 
all or most societies, is only permissible in ordinary social inter- 
course between persons standing in a specially familiar relationship. 
The prohibition against any reference to sexual matters before a 
father, and still more before a father-in-law, in many African 
societies, exemplifies this contrast between respectful and familiar 
or licentious behaviour. 

The theory, of which I gave a brief outUne in an earlier 
number of Africa^ and to which M. Griaule adopts a negative 
attitude, starts from the position that the customs of avoidance 
or extreme respect towards the wife's parents, and of privileged 
'joking' with the wife's brothers and sisters, can be regarded as the 
means of establishing and maintaining social equilibrium in a 
type of structural situation that results in many societies from 
marriage. In this situation we have two separate and distinct 
social groups, families or lineages, which are brought into con- 
nection with one another through the union of a man of one with 
a woman of the other. The husband is outside, and socially 
separated from, his wife's group. Through his relationship with 
her he is in an indirect or mediated relationship with individuals 
of her group. What is required for social equilibrium is that, as 
far as possible, he should not enter into conflict with his wife's 
group, but be obliged to maintain with that group or its members 
a 'friendly' relation. Both the avoidance customs and the 'joking' 
customs are the means by which this situation is socially regulated. 

Why the difference, then, between the behaviour towards 
the wife's parents and that towards her brothers and sisters? 
The answer lies in the general principle, widely recognised, that 
towards relatives of the first ascending generation respect is 
required, whereas relations of familiarity and equality are ap- 
propriate between persons of the same generation. There are, 
of course, examples of exceptions to this rule, such as joking 
relations or privileged familiarity tov/ards the father's sister's 
husband or the mother's brother. 

Thus the special structural situation considered in this theory 
is one of groups which maintain their separateness, each having 
its own system of internal relationships between its members, and 
an indirect connection of a person of one group with the other 

^ Africa, Vol. XIII, No. 3, 1940, pp. 195-210. See Chapter IV. 


group through a particular personal relation. In the instance of 
marriage the indirect relation is that of a man through his wife. 
The custom of a joking relationship with the mother's brother is ^ 
found in societies in which an individual belongs to a patrilineal f 
group, and therefore has an indirect relation to his mother's 1 
group through his mother. The widespread custom of relationships J 
of familiarity with grandparents, often taking the form of a joking 
relationship (in Australia, Africa, North America, the Oraons of 
India), emphasises the relation of the two generations as being 
socially separated. The grandparents are thus placed in contrast 
with the relatives of the parents' generation, and the relation to 
one's own grandparent is an indirect one through a parent. The J 
joking relationship between cross-cousins (Fiji, Ojibwa, etc.) 
is frequently a relation between possible relatives by marriage, 
but the relation is an indirect one through the mother or through 
the father's sister. 

An interesting crucial instance for the theory is provided 
by the Crow Indians, who have matrilineal clans. A man must 
be respectful to all the members of his father's clan; though he is 
not a member of that clan his relation to its members is one of 
close solidarity. In the other clans of the tribe there are to be found 
some men who are sons of men of his father's clan. They belong 
to clans that are separate and distinct, not only from his own clan 
but also from his father's clan. With such men, with whom his 
personal relation is an indirect one through his father's clan, 
he has a joking relationship; he may make offensive remarks to 
them or receive such from them without offence. In the Crow 
tribe this relationship has been developed into an instrument of 
social control of conduct, since the joking relative may call public 
attention to the shortcomings of his relative. 

The Cherokee also had a system of matrilineal clans, and a man 
was required to show respect to all the members of his father's clan. 
But with the clans of his father's father and his mother's father he 
had only an indirect connection through a parent. He called all 
the women of these clans 'grandmother' and could be on a 
relationship of familiarity or joking with them. Since marriage 
with such a 'grandmother' was approved they were possible 
wives or sisters-in-law. 

The theory that I have offered of joking relationships between 
persons related through marriage or by kinship is that they occur 


as social institutions in structural situations of a certain general 
kind in which there are two groups, the separateness of which is 
emphasised, and relations are established indirectly between a 
person in one group and the members or some of the members of 
the other. The relationship may be said to be one which expresses 
and emphasises both detachment (as belonging to separated 
groups) and attachment (through the indirect personal relation). 
These relationships of 'friendship', by avoidance or joking, 
contrast in a marked way with the relationships of solidarity, 
involving a complex system of obligations, that exist within a 
group such as a lineage or a clan. For the further development 
of the theory they need to be compared also with those relations 

-which are set up, between persons belonging to different groups, 
by the regular exchange of gifts. Thus the theory is only one part 
of an attempt to deal systematically with the types of social 
relationship that are to be found in primitive societies. 

^ The great majority of instances of joking relationships that 
were recorded by ethnographers were relationships between 
individuals connected through marriage or by kinship. Hence the 
reference to them in French as relations oiparenU. But there were 
also found instances of a similar relation between groups of persons, 
by which a member of one group was permitted and expected to 
offer insulting or derogatory remarks to any member of the other. 
A good example is provided by the 'coyote' and 'wild cat' 
moieties of Californian tribes. More recently similarcu stoms 
have been reported from Africa (Northern Rhodesia, Tanganyika, 
West Africa)^ in which this kind of relationship exists between two 
clans of a tribe or between two tribes. These obviously present 
a problem of a somewhat different kind. But it is obvious that 
any valid general theory of joking relationships must take into 
account these relations between groups. 

Tribes and clans are distinct separated groups each maintaining 
its own identity and separateness. Within a clan the relations of 
its members are those of solidarity in the special sense in which 
I have been using that term in this note. Two clans may, in some 
instances, be united in a way in which there is a permanent union 
of solidarity between the clans as groups and the members thereof. 
On the other hand there may also be a relation between two clans 

^See Bibliography, p. 115. 


of active or latent hostility. There is a third possibility, that 
between two particular clans there may be a relation neither of 
solidarity nor of hostility but of 'friendship' in which the separate- 
ness of the groups is emphasised, but open conflict between the 
groups or the members on the two sides is avoided by establishing 
a relation in which they may insult each other without giving 
or taking offence. This kind of thing is well illustrated in the 
account of the clans of the Tallensi given by Dr. Meyer Fortes. ^ 
A similar relationship, whereby hostility is avoided, may exist 
between two tribes, as in the instances known from Tanganyika.- 
It seems to me that in this way the joking relationship between 
clans and tribes recorded from Africa can be brought within the 
scope of a single theory that refers all instances of these relation- 
ships to a certain general type of structural situation. It should 
be made clear that what such a theory attempts to do is to deal 
with all the known examples of a certain recognisable type of 
institutionalised relationships in order to discover what common 
social feature makes this type of behaviour appropriate, meaning- 
ful, and functional. 

It is evident that in one particular respect the relation between 
the Dogon and the Bozo is similar to the relations that have been 
described from other parts of Africa, namely, in the exchange of 
insults. There is no evidence that they are similar in other 
respects, and they certainly are not so in all. The relation is 
spoken of as an 'alliance', but it is something very diflferent from 
an alliance between two nations which co-operate in fighting a 
war against another. The term 'alliance' is therefore not entirely 
suitable, nor have I been able to find a really suitable term. I 
have used the term 'friendship' and there is justification for this 
in the way in which native peoples themselves speak of friendship. 
In Australian tribes a man may have a 'friend',, that is, a person 
with whom he has a special personal relationship. In one region a 
wife's own sister's husband, if he is not a near kinsman, is such a 
friend. In other regions a man may not select a 'friend' from 
amongst the men to whom he applies the classificatory term for 
'brother'. Between 'brothers' relations are fixed by the kinship 
system. He may choose a man who stands to him in the 

^ Fortes, M., The Dynamics of Clanship among the Tallensi, London: Oxford 
University Press, 1945. 

^ See Bibliography, p. 115. 



classificatory relation of 'brother-in-law' but not his own brother- 
in-law. For brothers-in-law always necessarily belong to separated 
groups. There is a clear distinction made between friendship and 
relationships by kinship. 

I am distinguishing, therefore, a certain class of what I 
call 'friendship' relations, from what I have called relations of 
'solidarity' established by kinship or by membership of a group 
such as a lineage or clan. These terms are used only for the purpose s 
of the present analysis because in this matter, as in so many others 
in social anthropology, no precise technical terms are yet available. 

We may regard as one type of 'friendship' in this sense the 
relation set up between persons or groups on the basis of a con- 
tinued exchange of goods and services. The world-wide custom of 
gift-exchange has to be considered in this connection. But there 
are other varieties; one group may bury the dead of the other or 
perform other ritual services. In North-West America one group 
would call in a 'friend' group to erect a totem-pole for them. 
A component of the relationship between groups is very commonly 
a certain amount and kind of opposition, meaning by that term 
socially controlled and regulated antagonism. The two groups 
may regularly engage in competitive games such as football. In 
potlatch in North America there is competition or rivalry in 
exchange of valuables. Social relations of friendly rivalry are of 
considerable theoretical importance. The universities of Oxford 
and Cambridge maintain a certain relation by competing regu- 
larly in rowing, football, etc. The joking relationship is thus one 
example of a wider kind; for it is a relation of friendship in which 
there is an appearance of antagonism, controlled by conventional 

The 'alliance' between the Dogon and the Bozo described and 
analysed by M. Griaule is clearly an example of what I have been 
referring to as 'friendship'. The Dogon and the Bozo are separate 
peoples distinguished by language and by their mode of life. 
The prohibition against intermarriage maintains this separation by 
preventing the creation of relations of kinship between members 
of the two groups. The 'friendship' appears in the prohibition, 
under supernatural sanction, against the shedding of the blood 
of a member of the allied people, and in the regular exchange 
of gifts and services, for example, the services that individuals 
of one group perform in the ritual purification of those of the 


Other, To these is added a 'joking relationship', the exchange of 
insults between members of the two groups. It is with this last 
feature that we are concerned here. 

This alliance is conceived by the two peoples concerned in 
terms of their own cosmological system of myths and ideas, and 
M. Griaule's article is an important addition to the series of 
publications in which he and his co-workers have given the 
results of their investigation of this cosmology. It is in terms of 
these ideas that the Dogon interpret the exchange of insults.^ 
The exchange of insults is 'cathartic' because it rids the livers of 
both parties of impurities. M. Griaule has thus given us an ex- 
planation of the exchange of insults between Dogon and Bozo by 
showing what meaning it has to the natives themselves and also 
by showing its interconnections within a complex system of 
institutions, ideas, and myths. He finds that the most important 
function of the alliance is to provide what he calls, for lack of a 
more suitable term, 'purification'. So, provisionally, he proposes 
to call this type of alliance, as found in an extensive region of 
Africa, 'cathartic alliance'. Doubtless he would not suggest that we 
should apply this name to the exchange of insults between clans 
among the Tallensi or Bemba, or between tribes in Tanganyika. 
M. Marcel Mauss and I have both been seeking for many 
years to find a satisfactory general theory of what I have been 
calling relations of 'friendship' between separate groups of per- 
sons belonging to separate groups. One part of such a theory must 
be a study of prestations or exchanges of goods or services. 
Another must be a study of 'joking relationships'. It is towards 
such studies that M. Griaule adopts, as he says, 'a negative 
attitude'. He suggests that to classify together the various ex- 
amples of 'joking relationships' and to look for a general explana- 
tion, is like classifying together the ceremonies at which church 
bells are rung, such as funerals and weddings, calling them all 
ceremonies a cloches. This is the question of methodology in social 
anthropology that seems to me so important. For M. Griaule 
seems to be questioniiigLthe scientific validity of the com parative 
method as a means of arriving at general theoretical interpretations 
"of social institutions. 

It is only by the use of a comparative method that we can arrive 

1 Africa, Vol. XVIII, No. 4, pp. 253-4- 


at general explanations. The alternative is to confine ourselves 
to particularistic explanations similar to those of the historians. 
The two kinds of explanation are both legitimate and do not 
conflict; but both are needed for the understanding of societies 
and their institutions. That the Dogon explain the exchange of 
insults as a means of purifying the liver does not prevent us from 
treating the Dogon institution as one example of a very wide- 
spread form of 'friendship' in which such exchange is a distinctive 

It is not a question of whether my theory, or any other general 
theory, of joking relationships is or is not satisfactory. It is the 
different question of whether such a general theory is possible, 
or whether attempts to arrive at one should be abandoned in 
favour of resting content with particularistic explanations. 

The same question of methodology arises in connection with 
the conclusion of M. Griaule's article. He touches briefly on the 
need for an explanation of the Dogon-Bozo alliance 'en tant que 
systeme de groupes couples et dont les deux parties ont des 
prerogatives et devoirs complementaires'. He finds the explanation 
in 'les fondements meme de la metaphysique dogon. En effet, 
des I'origine du monde, la regie etait de gemelleite. Les etres 
devaient naitre par couple.' This is therefore a particularistic 
explanation in terms of Dogon ideas about twins. 

Relations of this kind between paired groups are to be found 
in many parts of the world. Outstanding examples are provided 
by the moiety organisations of North and South America, 
Melanesia, and Australia. The most usual way of representing 
this unity in duality, linking two groups into one society, is by 
pairs of opposites, such as heaven and earth, war and peace, red 
and white, land and water, coyote and wild cat, eaglehawk and 
crow. The underlying conception is therefore that of the union 
of opposites, as in the philosophy of Heraclitus. It was highly 
elaborated by the Chinese in the philosophy of Yin and Yang; 
yang and yin are male and female, day and night, summer and 
winter, activity and passivity, etc., and the dictum is that yang 
and yin together are required to make a unity or harmony {tao) 
as in the union of husband and wife, or the union of winter and 
summer to make a year. 

The Dogon are therefore unusual when they represent the 
relation between paired groups by reference to human twins. 


But this can be seen to be only a special development of a con- 
ception that is very widespread in Africa, by which twins are 
regarded as a single entity divided into two parts. A comparative 
study of African customs concerning twins shows this conception 
developed in a number of different ways. 

In the Dogon cosmology as recorded by M. Griaule and his 
associates the most fundamental conception of unity in duality 
seems to be not that of twin births but rather the opposition of 
the masculine and feminine principles, just as in the yin and yang 
of China. Human beings are born endowed with both principles 
and it is by operations of circumcision and clitoridectomy that 
they become truly male and female, so that there is again a 
Heraclitean union of opposites in the sexual union of husband 
and wife. One useful clue to the understanding of Dogon cos- 
mological ideas, or certain of them, is the way in which this 
duality of male and female is combined with the duality in unity 
of twins. The latter form of duality corresponds to the number 2; 
the former to the opposition between 3, masculine symbol, and 
4, feminine, which being added together give 7, the symbol of 
the complete being. 

The symbolic representations of the Dogon present striking 
similarities to those found in other parts of the world besides 
West Africa. The basis of any scientific understanding of them 
must be such a particularistic study as is being made by M. Griaule 
and his co-workers; but it is suggested that it will need to be 
supplemented by a systematic comparative study extended as 
widely as possible. The conception of unity in duality has been 
used by man not only in the establishment of systems of cosmology 
but also in organising social structures. A comparative study of 
this, as of joking relationships, may be expected to aid in most 
important ways the understanding of the Dogon system which, 
without it, would seem to be only a peculiar product of a particular 


Fortes, M. The Dynaynics of Clanship afuong the Talleiisi. London 

Oxford University Press, 1945. 
Moreau, R. E. 'The Joking Relationship (utani) in Tanganyika', 

Tanganyika Notes and Records, 12, 1941, pp. i-io. 
— 'Joking Relationships in Tanganyika', Africa, Vol. XIV, No. 3, 1944, 

pp. 386-400. 


Paulme, Denise. 'Parent^ k plaisanteries et alliance par le sang en Afrique 

occidentale', Africa, Vol. XII, No. 4, 1939, pp. 433-44. 
Pedler, F. J. 'Joking Relationships in East Africa', Africa, Vol. XIII, No. 

2, 1940, pp. 170-3. 
Radcliffe-Brown, A. R. 'On Joking Relationships', Africa, Vol. XII, 

No. 3, 1940, pp. 195-210. 
Richards, A. I. 'Reciprocal Clan Relationships among the Bemba of 

N. Rhodesia', Man, Vol. XXXVII, p. 222, 1927. 
Schapera, I. 'Customs relating to twins in South Ahica\ yournal Afr. Soc, 

Vol. XXVI, cii, pp. 117-37. 



THERE has been in the past some disagreement and discussion 
as to the definition of totemism. I wish to avoid as far as 
possible entering into any such discussion. The purpose 
of preUminary definitions in science is to mark off a class of 
phenomena for special study. A term is useful if and in so far as 
it brings together for our attention a number of phenomena which 
are in reality, and not merely in appearance, closely related to 
one another. It will be part of my thesis in this paper that 
however widely or narrowly we may define totemism, we cannot 
reach an understanding of the phenomena we so name unless we 
study systematically a much wider group of phenomena, namely, 
the general relation between man and natural species in mytholog}' 
and ritual. It may well be asked if 'totemism' as a technical term 
has not outlived its usefulness. . . - 

It is necessary, however, to have some definition to guide and 
control our discussion. I shall use the term in the wider sense to 
apply wherever a society is divided into groups and there is a 
special relation between each group and one or more classes of 
objects that are usually natural species of animals or plants but 
may occasionally be artificial objects or parts of an animal. The 
word is sometimes used in a narrower sense and applied only 
when the groups in question are clans, i.e. exogamous groups of 
which all the members are regarded as being closely related by 
descent in one line. I shall regard 'clan totemism' as only one 

variety of totemism, in the wider sense. ^ 

Even in the narrower sense of clan totemism, and still more 
in the wider sense, totemism is not one thing but is a general name 

^ Reprinted from Proceedwgs of the Fourth Pacific Science Congress, Ja^'a, 

^ It is sometimes said that totemism has two aspects, a social aspect and a 
religious or ritual aspect. What is referred to as the 'social aspect' of totemism 
is simply the clan organisation. But exogamous clans similar in all essentials to 
totemic clans so far as economic or juridical functions go, can, as we well know, 
exist without totemism. The so-called 'social aspect' of clan totemism is simply 
the social aspect of the clan. 



given to a number of diverse institutions which all have, or seem 
to have, something in common. Thus even in the Hmited region 
of Austraha, which has a single homogeneous culture throughout, 
there have been recorded a number of different varieties of 
totemism, and new varieties are being discovered by systematic 
researches now in progress. 

In the south-east of the continent is found sex-totemism, i.e. 
an association of the two sex-groups, men and women, with two 
animal species. In the coastal districts of New South Wales, for 
example, the bat is the totem or animal representative of the men 
and the tree-creeper {Climacteris sp.) is that of the women. 

In many parts of Australia the tribe is divided into two 
exogamous moieties, patrilineal in some regions, matrilineal in 
others. In some instances thte moieties are named after species of 
animals, generally birds. Amongst such names are the following 
pairs: crow and white cockatoo, white cockatoo and black cockatoo, 
eaglehawk and crow, native companion and turkey, hill kangaroo 
and long-legged kangaroo. In other instances the meanings of the 
moiety names have not been discovered, and in some of them, at 
any rate, it seems certain that they are not animal names. 

In many of the tribes that have this dual division, indepen- 
dently of whether the moieties are named after animals or not, 
there is a classification of animals and frequently of other natural 
objects whereby some are regarded as belonging to one moiety 
and others to the other. 

Such moiety totemism, if we may use that term for any such 
association between the moiety and one or more natural species, 
is found in a number of different varieties in Australia, and still 
other varieties are found in Melanesia and in North America. 

Over a large part of Australia the tribe is divided into four 
groups which have often been called 'classes' but which I prefer 
to call 'sections'. The easiest way to understand this division 
into four is to regard it as constituted by the intersection of a pair 
of patrilineal moieties and a pair of matrilineal moieties.^ 

^ If we denote the four sections as A, B, C, and D, the matrilineal moieties 
are A + C and B + D; the patrilineal moieties are A + D and B + C. Since 
a man may not marry within his o^-n patrilineal moiety or within his own 
matrilineal moiety it will follow that a man of A can only marry a woman of B 
and their children must belong to section D, i.e. to the patrilineal moiety of the 
father (A) and to the matrilineal moiety of the mother (B). 


These sections are not as a rule named after species of animals, 
though there are one or two instances in which a section name is 
also the name of an animal. Thus Bandjur in Yukumbil is the 
name of a section and also of the native bear. In some tribes, 
however, there is a definite association between each section and 
one or more species of animal. Thus in the Nigena tribe of the 
Kimberley district of Western Australia the four sections are 
associated with four species of hawk. In some regions this associa- 
tion does not carry with it any prohibition against killing or eating 
the animal associated with one's own or any other section. In 
part of Queensland, however, each section has associated with it a 
number of species of animals and there is a rule that the members 
of a section may not eat the animals so associated with their 

This 'section totemism' requires further investigation. We 
may distinguish, however, three varieties. In one each section has 
associated with it a single species of animal which is representative 
of the section in somewhat the same way as the sex-totem is the 
representative of the sex-group. In a second variety each section 
stands in a special ritual relation to a certain limited number of 
species which may not be eaten by the members of the section. 
In the third variety a great number of species of animals are classi- 
fied as belonging to one or other of the four sections but there is no 
rule against eating the animals belonging to one's own section. 
The one thing that is common to these varieties is that each 
section is differentiated from the others and given its own in- 
dividuality by being associated with one or more animal species. 

In some tribes the four sections are again subdivided each into 
two parts, giving a division of the tribe into eight sub-sections. 
In some of these tribes there exist special associations between 
the sub-sections and certain natural species. Further investigation 
is needed before we can profitably discuss this subject. 

If now we turn to clan-totemism we find a number of different 
varieties of this in Australia, too many, in fact, to be even enu- 
merated in a short paper. Matrilineal clan totemism of different 
varieties occurs in three, or possibly four, separate areas in the 
east, north, and west of the continent. In Melville and Bathurst 
Islands there are three matrilineal phratries subdivided into 
twenty-two clans. Each clan is associated with one natural species, 
usually a species of animal or plant, though one or two clans have . 


two totems and one has three. The association between the clan 
and its totem is aipparently of very Httle importance in the Hfe 
of the tribe. There is no prohibition against eating or using the 
totem, there are no totemic ceremonies, and totemism has httle 
influence on the mythology. 

The matrilincal clan totemism of some tribes of New South 
Wales, Victoria and South Australia seems to be of somewhat 
more importance. Here we find matrilineal moieties sometimes 
named totemically, sometimes not, and each moiety is divided into 
a number of clans. Each clan has one or more natural species 
regarded as belonging to it. Where there are several species 
associated with each clan, as is the case in many tribes, one of 
them is regarded as more important than the others and the clan 
is named after it. Throughout this region there is, so far as we 
know, no prohibition against killing or eating the totem. 

Totemic ceremonial is apparently little developed nor have we 
any evidence of any elaborate totemic mythology connected with 
matrilineal totemism. 

It should be noted that throughout Australia the most im- 
portant group for social purposes is the horde, i.e. a small group 
occupying and ow^ning a certain defined territory, and that the 
horde is normally strictly patrilineal. It follows that wherever 
there is a system of matrilineal totemic clans the clan consists of 
individuals scattered through a number of hordes. We thus get a 
double grouping of individuals. For most social purposes the 
individual is dependent on the local group, i.e. the horde, to which 
he is connected through his father, while at the same time he is 
also connected through his mother to a totemic group the members 
of which are scattered throughout the tribe. 

Patrilineal totemism in Australia is more difficult to describe 
briefly than is matrilineal totemism. Where it exists the primary 
totemic group is usually the horde, i.e. the small patrilineal local 
group. In some regions the horde is a clan, i.e. it consists of close 
relatives in the male line and is therefore exogamous. But in a 
few regions the horde is not a clan in this sense. 

As an example of one variety of patrilineal totemism we may 
take the tribes at the mouth of the Murray River (Yaralde, etc.). 
Here each horde is a local clan and each clan has one or more 
species of natural object associated w ith it. There is no prohibition 
against eating the totem of one's clan, but it is regarded with some 


respect. There is no evidence of totemic ceremonial or of any 
elaborate totemic mythology. The function of the totem seems to 
be merely to act as the representative of the group. 

Perhaps the most important, and certainly the most interesting, 
form of totemism in Australia is that to a brief consideration of 
which we now pass. This consists of a fourfold association between 
(i) the horde, i.e. the patrilineal local group, (2) a certain number 
of classes of objects, animals, plants, and other things such as rain, 
sun and hot weather, cold weather, babies, etc., (3) certain sacred 
spots within the territory of the horde, frequently water-holes, 
each one of which is specially associated with one or more of the 
'totems' of the group, and (4) certain mythical beings who are 
supposed to have given rise to these sacred spots in the mythical 
period of the beginning of the world. This system of totemism 
is now being traced and studied in a number of variant forms over 
a very large part of the Australian continent. It \^"as formerly 
known best from the centre of the continent, where, however, the 
Aranda have it in a somewhat modified or anomalous form. 
We now know that it exists or existed over a large part of Western 
Australia. Recently it has been discovered and studied in the 
Cape York Peninsula by Miss McConnel. At the beginning of this 
year I was able to demonstrate its former existence on the east 
coast of Australia in the north of New South Wales and in southern 

Where this type of totemism is found it is usually accompanied 
by a system of ceremonies for the increase of natural species. The 
members of the horde, or some of them, proceed to the totem 
centre or sacred spot connected with a natural species and perform 
there a ceremony which is believed to result in an increase of 
that species. There is also an elaborate mythology dealing with the 
sacred totem centres and with the mythical beings who gave rise 
to them. 

It may be noted that this kind of totemism may coexist in 
the same tribe with other kinds. Thus in the Dieri tribe it exists 
together with a system of matrilineal clan totemism. In some 
parts it coexists v.ith section totemism. 

Finally, we may note that in some parts of Australia there 
exists what is sometimes called individual or personal totemism. 
This is a special relation between an individual and some one or 
more species of animal. A good example is found in some tribes 


of New South Wales where every medicine-man has one or more 
of such personal totems. It is through his association with the 
animal species that he acquires his power to perform magic. 
Whether we call this totemism or not, it is quite evident that it is 
closely related to totemism and that any theory of totemism, to be 
satisfactory, must take it into account. 

This brief and very incomplete survey of Australian institutions 
has shown us that special associations of groups or individuals 
with natural species exist in that region in a number of different 
forms. We find all gradations from a tribe with no form of totemism 
at all (such as the Bad of northern Dampier Land) through tribes 
such as the Melville Islanders where totemism of a simple form 
exists but is of comparatively little importance in the life of the 
tribe, to a tribe such as the Dieri which combines in a complex 
system two forms of totemism, one of matrilineal clans and the 
other of patrilineal hordes, with a highly elaborated totemic 
ritual and mythology. The only thing that these totemic systems 
have in common is the general tendency to characterise the seg- 
ments into which society is divided by an association between 
each segment and some natural species or some portion of nature. 
The association may take any one of a number of different forms. 

In the past the theoretical discussion of totemism was almost 
entirely concerned with speculations as to its possible origin. 
If we use the word origin to mean the historical process by which 
an institution or custom or a state of culture comes into existence, 
then it is clear that the very diverse forms of totemism that exist 
all over the world must have had very diverse origins. To be 
able to speak of an origin of totemism we must assume that all 
these diverse institutions that we include under the one general 
term have been derived by successive modifications from a single 
form. There does not seem to me to be a particle of evidence to 
justify such an assumption. But even if we make it we can still 
only speculate as to what this original form of totemism may have 
been, as to the enormously complex series of events which could 
have produced from it the various existing totemic systems, and 
as to where, when, and how that hypothetical original form of 
totemism came into existence. And such speculations, being for 
ever incapable of inductive verification, can be nothing more 
than speculations and can have no value for a science of culture. 

For sociology, or social anthropology, by which I understand 



the Study of the phenomena of culture by the same inductive 
methods that are in use in the natural sciences, the phenomena of 
totemism present a problem of a different kind. The task of the 
inductive sciences is to discover the universal or the general in the 
particular. That of a science of culture is to reduce the complex 
data with which it deals to a limited number of general laws or 
principles. Approaching totemism in this way we may formulate 
the problem that it presents in the form of the question, 'Can we 
show that totemism is a special form of a phenomenon which is 
universal in human society and is therefore present in different 
forms in all cultures?' 

The most important attempt to arrive at a sociological theory 
of totemism is that of the late Professor Durkheim in his work 
Les Formes elementaires de la Vie religieuse. I think that that work 
is an important and permanent contribution to sociological theory, 
but that it does not provide a complete and satisfactory theory 
of totemism. I shall attempt to point out, in the briefest possible 
way, where Durkheim's theor)^ seems to me to fail. 

Durkheim speaks of the totem as being 'sacred' to the members 
of the group of which it is the totem. This is to use the term 
'sacred' in a sense somewhat different from that which it has at the 
present day in English or even in French, and not even identical 
with, though somewhat nearer to, the meaning that sacer had in 
Latin. I prefer to use a term which is as free as possible from 
special connotations, and therefore instead of saying that the 
totem is sacred I find it preferable to say that there is a 'ritual 
relation' between persons and their totem. There exists a ritual 
relation whenever a society imposes on its members a certain 
attitude towards an object, which attitude involves some measure 
of respect expressed in a traditional mode of behaviour with 
reference to that object. Thus the relation between a Christian 
and the first day of the week is a typical example of a ritual 

Every society adopts, and imposes upon its members, to\\ards 
certain objects, this attitude of mind and behaviour which I am 
calling the ritual attitude. There are, not only in different societies, 
but in the same society in different references, many different 
varieties of this attitude, but all the varieties have something in 
common. Moreover the ritual attitude may vary from a very 
indefinite one to a definite and highly organised one. 


One of the important problems of sociology is therefore to 
discover the function of this universal element of culture and 
to formulate its laws. This general problem obviously includes a 
vast number of partial problems of which the problem of totemism 
is one. That problem may be stated as being that of discovering 
why in certain societies a ritual attitude towards a certain species 
of natural object is imposed upon the members of a particular 
social group. It is obvious that no solution of the lesser problem 
of totemism can be satisfactory unless it conforms with or is part 
of a general solution of the wider problem, i.e. a theory of ritual 
relations in general. 

With regard to the general problem Durkheim's theory is that 
the primary object of the ritual attitude is the social order itself, 
and that any thing becomes an object of that attitude when it 
stands in a certain relation to the social order. This general 
theory, with which I agree, obviously amounts to very little until 
we have succeeded in defining the more important types of 
relation to the social order which result in the object which stands 
in such a relation becoming an object of ritual attitude. 

If I may restate in my own terms Durkheim's theory of totem- 
ism it is as follows. A social group such as a clan can only possess 
solidarity and permanence if it is 'the object of sentiments of 
attachment in the minds of its members. For such sentiments 
to be maintained in existence they must be given occasional 
collective expression. By a law that can be, I think, readily 
verified, all regular collective expressions of social sentiments 
tend to take on a ritual form. And in ritual, again by a necessary 
law, some more or less concrete object is required which can act 
as the representative of the group. So that it is a normal procedure 
that the sentiment of attachment to a group shall be expressed 
in some formalised collective behaviour having reference to an 
object that represents the group itself. 

A typical example is to be found in our own society. National 
solidarity depends on a sentiment of patriotism in the minds of the 
nation. This sentiment, in conformity with the laws stated above, 
tends to find some of its chief expressions in reference to such 
concrete objects as flags, or kings and presidents, and such objects 
become in this way objects of the ritual attitude. 

Part of a king's sacredness, whether in Africa or in Europe, 
is due to the fact that he is the representative of the national 


solidarity and unity, and the ritual that surrounds him is the 
means by which patriotic sentiments are maintained. In the same 
way in the flag we have an object which is 'sacred' because it is 
the concrete material representative or emblem of a social group 
and its solidarity. 

Durkheim compares the totem of a clan with the flag of a 
nation. The comparison is valid, in a very general sense, for some 
forms of totemism, if not for all. But putting the comparison aside, 
the theory is that the totem is 'sacred' as Durkheim says, or is an 
object of ritual attitude, as I prefer to say, because it is the concrete 
representative or emblem of a social group. And the function of 
the ritual attitude towards the totem is to express and so to main- 
tain in existence the solidarity of the social group. 

With Durkheim's theory as stated above in my own terms I 
am in agreement, but I do not regard it as complete. In the first 
place it seems to me that totemism has other functions besides the 
one indicated above. Secondly, the theory so far as stated above 
does not explain why so many peoples in America, Asia, Africa and 
Australasia should select as emblems or representatives of clans 
or other social groups species of animals or plants. It is true that 
Durkheim off"ers an answer to this question, but it is an entirely 
unsatisfactory one. He regards as an essential part of totemism the 
use of totemic emblems or designs, i.e. figured representations 
of the totemic animal or plant, and suggests that the reason for 
selecting natural objects as emblems of social groups is because 
they are capable of being used in this way. 

This hypothesis fails as soon as we apply it to the facts. In 
Australia no designs are made of the sex totems or of the totems of 
the moieties or sections, and even for clan totemism there are many 
tribes that do not make any representations of their totems. 
Totemic designs, which for Durkheim are so important or indeed 
so essential a part of totemism, are characteristic of central and 
northern Australia but not of the continent as a whole. 

Moreover, the reason suggested for the selection of natural 
objects as emblems of social groups is of too accidental a character 
to give a satisfying explanation of an institution that is so wide- 
spread as totemism. There must surely be some much more im- 
portant reason why all these peoples all over the world find it 
appropriate to represent social groups in this way by associating 
each one with some animal or plant. 


This, then, is where I think Durklicim's theory' of totemism 
fails. It implies that the totem owes its sacred or ritual character 
solely to its position as the emblem of a group. Now there are a 
number of peoples who have no form of totemism amongst whom 
we still find that natural species such as animals and plants are 
objects of ritual or of the ritual attitude expressed in mythology. 
And even amongst totemic peoples such as the Australian tribes 
the ritual customs relating to natural species are not all totemic. 
In other words the phenomena which we have agreed to denote 
by the term totemism are merely a part of a much larger class of 
phenomena which includes all sorts of ritual relations between 
man and natural species. No theory of totemism is satisfactory 
unless it conforms with a more general theory providing an 
explanation of many other things besides totemism. Durkheim's 
theory fails to do this. 

In a great number, and I believe probably in all, of the societies 
where man depends entirely or largely on the hunting of wild 
animals and the collection of wild plants, whether they have any 
form of totemism or not, the animals and plants are made objects 
of the ritual attitude. This is done frequently, though perhaps not 
quite universally, in mythology, in which animal species are 
personified and regarded as ancestors or culture heroes. It is done 
also by a mass of customs relating to animals and plants. This 
system of ritual and mythological relations between man and 
natural species can be best studied in non-totemic peoples such 
as the Eskimo or the Andaman Islanders. In such societies we 
find that the relation between the society and the natural species 
is a general one, all the most important animals and plants being 
treated as in some way sacred (either in ritual or in mythology) 
and some being regarded as more sacred than others, but any 
single species being equally sacred to every member of the whole 
community. The ritual attitude of the Andaman Islanders towards 
the turtle, of Californian Indians to the salmon, of the peoples of 
North America and northern Asia to the bear, constitutes a 
relation between the whole society and the sacred species. 

Now totemism, I would suggest, arises from or is a special 
development of this general ritual relation between man and 
natural species. Let us assume for the moment that such a general 
ritual relation of man to nature is universal in hunting societies, 
as I believe it can be shown to be. When the society becomes 


differentiated into segmentary groups such as clans, a process of 
ritual specialisation takes place by which each segment acquires a 
special and particular relation to some one or more of the sacra 
of the community, i.e. to some one or more natural species. 
The totem of the clan or group is still sacred in some sense to 
the whole community, but is now specially sacred, and in some 
special way, to the segment of which it is the totem. 

The process here suggested as the active principle in the 
development of totemism is one which I believe to be of great 
importance in social development, and which can be observed in 
other phenomena. Thus, to take only one example, and perhaps 
not the best, in the Roman Church the saints are sacred to all 
members of the church as a whole. But the church is segmented 
into local congregations and a congregation is often placed in a 
special relation to one particular saint to whom its chapel is 
dedicated. This is, I think, parallel to clan or group totemism. 
We might also draw a significant though not quite exact analogy 
between the patron saint of an individual and the personal totem 
or guardian animal of Australian and American tribes. 

There is no space in this paper to discuss this process of 
ritual specialisation, and indeed any adequate treatment of the 
subject would require us to deal with the vvhole process of social 
differentiation and segmentation. I will refer to a single example 
that may help to illustrate the problem. Amongst the Eskimo of 
part of North America one of the most important features of their 
adaptation to their environment is the sharp division between 
winter and summer, and between the winter animals and the 
summer animals. There is a complex system of ritual relations 
between the society and all the most important of these animals 
and in this ritual the opposition between summer and winter 
is strongly expressed. Thus you may not eat reindeer meat 
, (summer food) and walrus meat (winter food) on the same day. 
The Eskimo have made for themselves a segmentation into two 
groups, one consisting of all the persons born in the summer and 
the other of those born in the \\intcr, and there is some slight 
ritual specialisation, the summer people being regarded as specially 
connected with the summer animals and the winter people with the 
winter animals. This is not quite totemism, but it is clearly related 
to it, and illustrates, I think, the process by which totemism arises. 

In this way, I think, we can formulate a sociological theory 


of totemism which incorporates a great deal of Durkheim's 
analysis and is not open to the criticisms that can be levelled 
against Durkheim's own presentation. We start with the em- 
pirical generalisation that amongst hunting and collecting peoples 
the more important animals and plants and natural phenomena 
are treated, in custom and in myth, as being 'sacred', i.e. they 
are made, in various ways and in different degrees, objects of the 
ritual attitude. Primarily this ritual relation between man and 
nature is a general one between the society as a whole and its 
sacra. When the society is differentiated, i.e. divided into seg- 
ments or social groups marked off from one another and each 
having its own solidarity and individuality, there comes into action 
a principle which is more widespread than totemism and is 
indeed an important part of the general process of social differ- 
entiation, a principle by which within the general relation of the 
society to its sacra there are established special relations between 
each group or segment and some one or more of those sacra. 

This theory incorporates what I think is the most valuable 
part of Durkheim's analysis, in the recognition that the function of 
the ritual relation of the group to its totem is to express and so to 
maintain in existence the solidarity of the group. It gives moreover 
a reason, which can be shown, I think, to be grounded in the 
very nature of social organisation itself, for the selection of natural 
species as emblems or representatives of social groups. 

Before leaving this part of the discussion I would like to touch 
on one further point. Durkheim, in reference to clan totemism, 
emphasises the clan and its solidarity. The totem, for him, is 
primarily the means by which the clan recognises and expresses 
its unity. But the matter is much more complex than this. The 
clan is merely a segment of a larger society which also has its 
solidarity. By its special relation to its totem or totems the clan 
recognises its unity and its individuality. This is simply a special 
example of the universal process by which solidarity is created 
and maintained by uniting a number of individuals in a collective 
relation to the same sacred object or objects. By the fact that each 
clan has its own totem there is expressed the differentiation and 
opposition between clan and clan. The kangaroo men not only 
recognise the bond that unites them as kangaroo men but also 
recognise their difference from the emu men and the bandicoot 
men and so on. But also the wider unity and solidarity of the 


whole totemic society is expressed by the fact that the society as a 
whole, through its segments, stands in a ritual relation to nature 
as a whole. This is seen very well in the system of increase 
ceremonies that is so widespread in Australia. Each group is 
responsible for the ritual care of a certain number of species by 
which the maintenance of that species is believed to be assured. 
For the tribe all these species are of importance, and the ceremonies 
are thus a sort of co-operative effort, involving a division of 
(ritual) labour, by which the normal processes of nature and the 
supply of food are provided for. One of the results of Durkheim's 
theory is that it over-emphasises the clan and clan solidarity. 
Totemism does more than express the unity of the clan; it also 
expresses the unity of totemic society as a whole in the relations 
of the clans to one another within that wider unity. 

The result of my argument, if it is valid, is to substitute for 
the problem of totemism another problem. The question that now 
demands an answer is, 'Why do the majority of what are called 
primitive peoples adopt in their custom and myth a ritual attitude 
towards animals and other natural species?' My aim in this paper 
has simply been to exhibit as exactly as possible in a brief space 
the relation of the problem of totemism to this wider problem. 

It is obvious that I cannot attempt in a mere conclusion to a 
paper to deal with this subject of the relation in myth and ritual 
of man and nature. I attempted some years ago to deal with it in 
reference to the customs and beliefs of one non-totemic people, the 
Andaman Islanders. As a result of that and other investigations I 
was led to formulate the following law: Any object or event which 
has important effects upon the well-being (material or spiritual) 
of a society, or any thing which stands for or represents any such 
object or event, tends to become an object of the ritual attitude. 

I have given reasons for rejecting Durkheim's theory that in 
totemism natural species become sacred because they are selected 
as representatives of social groups, and I hold, on the contrar}', 
that natural species are selected as representatives of social 
groups, such as clans, because they are already objects of the 
ritual attitude on quite another basis, by virtue of the general 
law of the ritual expression of social values stated above. 

In modern thought we are accustomed to draw a distinction 
between the social order and the natural order. We regard society 
as consisting of certain human beings grouped in a social structure 


under certain moral principles or laws, and we place over against 
the society its environment, consisting of geographical features, 
flora and fauna, climate with its seasonal changes, and so on, 
governed by natural law. 

For certain purposes this contrast of society and environment, 
of man and nature, is a useful one, but we must not let it mislead 
us. From another and very important point of view the natural 
order enters into and becomes part of the social order. The 
seasonal changes that control the rhythm of social life, the animals 
and plants that are used for food or other purposes, these enter 
into and become an essential part of the social life, the social order. 
I believe that it can be shown that it is just in so far as they thus 
enter into the social order that natural phenomena and natural 
objects become, either in themselves, or through things or beings 
that represent them, objects of the ritual attitude, and I have 
already tried to demonstrate this so far as the Andaman Islanders 
are concerned. Our own explicit conception of a natural order 
and of natural law does not exist amongst the more primitive 
peoples, though the germs out of which it develops do exist in the 
empirical control of causal processes in technical activities. For 
primitive man the universe as a whole is a moral or social order 
governed not by what we call natural law but rather by what 
we must call moral or ritual law. The recognition of this concep- 
tion, implicit but not explicit, in ritual and in myth is, I believe, 
one of the most important steps towards the proper understanding 
not only of what is sometimes called 'primitive mentality' but 
also of all the phenomena that w^e group vaguely around the 
term religion.^ 

A study of primitive myth and ritual from this point of view 
is, I think, very illuminating. In Australia, for example, there are 
innumerable ways in which the natives have built up between 
themselves and the phenomena of nature a system of relations 
which are essentially similar to the relations that they have built 
up in their social structure between one human being and another. 

^ A more precise way of stating the view I am here suggesting is that in every 
human society there inevitably exist two different and in a certain sense con- 
flicting conceptions of nature. One of them, the naturalistic, is implicit every- 
where in technology, and in our twentieth century European culture, with its 
great development of control over natural phenomena, has become explicit and 
preponderant in our thought. The other, which might be called the mythological 
or spiritualistic conception, is implicit in myth and in religion, and often 
becomes explicit in philosophy. 


I can do no more than mention examples. One is the personi- 
fication of natural phenomena and of natural species. A species of 
animal is personified, i.e. treated for certain purposes as if it were 
a human being, and in the mythology such personified species 
are regarded as ancestors or culture heroes. The function of this 
process of personification is that it permits nature to be thought 
of as if it were a society of persons, and so makes of it a social or 
moral order. Another of the processes by which, in Australia, the 
world of nature is brought within the social order is to be found 
in the systems of classification of natural species, existing in a 
number of diverse forms in different parts of the continent with 
this one thing in common to them all, that the more important 
natural species are so classified that each one is regarded as 
belonging to a certain social group, and occupying a specific 
position in the social structure. 

Although there is always a danger in short formulas I think. 
it does not misrepresent Australian totemism to describe it as a 
mechanism by which a system of social solidarities is established 
between man and nature. The mechanism has been worked out in 
many different ways, and much more elaborately in some than in 
others, but everywhere it possesses this character. 

The suggestion I put forward, therefore, is that totemism is 
part of a larger whole, and that one important w^ay in which we can 
characterise this whole is that it provides a representation of the 
universe as a moral or social order. Durkheim, if he did not actually 
formulate this view, at any rate came near to it. But his conception 
seems to have been that the process by which this takes place 
is by a projection of society into external nature. On the contrary, 
I hold that the process is one by which, in the fashioning of 
culture, external nature, so called, comes to be incorporated in 
the social order as an essential part of it. 

Now the conception of the universe as a moral order is not 
confined to primitive peoples, but is an essential part of every 
system of religion. It is, I think, a universal element in human 
culture. With the question of why this should be so I cannot now 
attempt to deal. 

I may summarise what I have tried to say as follows: A 
sociological theory of totemism must be able to show that totemism 
is simply a special form taken in certain definite conditions by an 
element or process of culture that is universal and necessary. 



Durkheim's attempt to provide such a theory fails in certaii 
important respects. We can, however, incorporate a good dea 
of Durkheim's analysis in a theory which rests on the sam( 
general hypothesis of the nature and function of ritual or th< 

Finally, my argument has brought out something of the con 
ditions in which this universal element of culture is most likeb 
to take the form of totemism. These are (i) dependence wholh 
or in part on natural productions for subsistence, and (2) th( 
existence of a segmentary organisation into clans and moieties o: 
other similar social units. The Andamanese and the Eskimo hav< 
(i) but not (2), and they have no totemism though they have th( 
material out of which totemism could easily be made. There are 
of course, apparent exceptions to this generalisation, in som< 
of the tribes of Africa, America and Melanesia. The detailec 
examination of these, which of course cannot be undertaken in i 
brief paper, really serves, I believe, to confirm the rule. 

I would not be understood to maintain the view that totemism 
or rather the diflferent institutions which in different parts of th< 
world we call by this general term, have arisen independently o 
one another. I think that it is very likely. But it does not matter fo: 
the sociologist, at any rate in the present state of our knowledge 
If anyone wishes to believe that all the existing forms of totemisn: 
have come into existence by a process of what is rather unsatisfac- 
torily called 'diffusion' from a single centre, I have no objection 
I would point out that totemism has not spread everywhere, oi 
evenly, and that it has not survived equally in all regions. It is 
sufficient for my argument if we can say that it is only where ^sr 
certain other features of culture are present that totemism is 
likely to be accepted by a people when it is brought to them from 
outside, or is likely to remain in active existence after it has 
been introduced. 







THE purpose of this lecture, which you have done me the 
honour of inviting me to deHver, is to commemorate 
the work of Sir James Frazer, as an example of life-long 
Qfbingle-minded devotion to scientific investigation and as having 
vepontributed, in as large a measure as that of any man, to laying 
(ifthe foundations of the science of social anthropology. It therefore 
eieems to me appropriate to select as the subject of my discourse 
pne which Sir James was the first to investigate systematically 
(lalf a century ago, when he wrote the article on 'Taboo' for the 
ijiinth edition of the Encyclopaedia Britannica, and to the elucida- 
;ion of which he has made many successive contributions in his 
vritings since that time. 

The English word 'taboo' is derived from the Polynesian 
vord 'tabu' (with the accent on the first syllable). In the languages 
)f Polynesia the word means simply 'to forbid', 'forbidden', and 
:an be applied to any sort of prohibition. A rule of etiquette, an 
)rder issued by a chief, an injunction to children not to meddle 
vith the possessions of their elders, may all be expressed by the 
ise of the word tabu. 

The early voyagers in Polynesia adopted the word to refer to 
prohibitions of a special kind, which may be illustrated by an 
xample. Certain things such as a newly-born infant, a corpse or 
he personof a chief are said to be tabu. This means that one should, 
Jis far as possible, avoid touching them. A man who does touch 
3i)ne of these tabu objects immediately becomes tabu himself. 
This means two things. In the first place a man who is tabu 
n this sense must observe a number of special restrictions on his 
jehaviour; for example, he may not use his hands to feed himself, 
^e is regarded as being in a state of danger, and this is generally 
tated by saying that if he fails to observe the customary pre- 
:autions he will be ill and perhaps die. In the second place he is 
Iso dangerous to other persons — he is tabu in the same sense 
s the thing he has touched. If he should come in contact with 

^ The Frazer Lecture, 1939. 


Utensils in which, or the fire at which, food is cooked, the dangerous 
influence would be communicated to the food and so injure 
anyone who partook of it. A person who is tabu in this way, as 
by touching a corpse, can be restored to his normal condition 
by rites of purification or desacralisation. He is then said to be 
Jioa again, this term being the contrary of tabu. 

Sir James Frazer has told us that when he took up the study 
of taboo in 1886 the current view of anthropologists at the time 
was that the institution in question was confined to the brown and 
black races of the Pacific, but that as a result of his investigations 
he came to the conclusion that the Polynesian body of practices 
and beliefs *is only one of a number of similar systems of super- 
stition which among many, perhaps all the races of men have 
contributed in large measure, under many different names and 
with many variations of detail, to build up the complex fabric 
of society in all the various sides or elements of it which we 
describe as religious, social, political, moral and economic'. 

The use of the word taboo in anthropology for customs all 
over the world which resemble in essentials the example given 
from Polynesia seems to me undesirable and inconvenient.' 
There is the fact already mentioned that in the Polynesian 
language the word tabu has a much wider meaning, equivalent, 
to our own word 'forbidden'. This has produced a good deal of, 
confusion in the literature relating to Polynesia owing to the 
ambiguity resulting from two different uses of the same word.' 
You will have noticed that I have used the word taboo (with the 
English spelling and pronunciation) in the meaning that it has fori 
anthropologists, and tabu (with the Polynesian spelling and, 
pronunciation) in special reference to Polynesia and in the Poly- 
nesian sense. But this is not entirely satisfactory. 

I propose to refer to the customs we are considering as 'ritual 
avoidances' or 'ritual prohibitions' and to define them by refer-! 
ence to two fundamental concepts for which I have been inj 
the habit of using the terms 'ritual status' and 'ritual value'.! 
I am not suggesting that these are the best terms to be found; 
they are merely the best that I have been able to find up to the 
present. In such a science as ours words are the instrument 
of analysis and we should always be prepared to discard inferio: 
tools for superior when opportunity arises. 

A ritual prohibition is a rule of behaviour which is associate 


TABOO 135 

with a belief that an infraction will result in an undesirable change 
in the ritual status of the person who fails to keep to the rule. 
This change of ritual status is conceived in many different ways 
in different societies, but everywhere there is the idea that it 
involves the likelihood of some minor or major misfortune which 
will befall the person concerned. 

We have already considered one example. The Polynesian 
who touches a corpse has, according to Polynesian belief, under- 
gone what I am calling an undesirable change of ritual status. 
The misfortune of which he is considered to be in danger is 
illness, and he therefore takes precautions and goes through a 
ritual in order that he may escape the danger and be restored to 
his former ritual status. 

Let us consider two examples of different kinds from con- 
temporary England. There are some people who think that one 
should avoid spilling salt. The person who spills salt will have 
bad luck. But he can avoid this by throwing a pinch of the spilled 
salt over his shoulder. Putting this in my terminology it can be 
said that spiUing salt produces an undesirable change in the ritual 
status of the person who does so, and that he is restored to his 
normal or previous ritual status by the positive rite of throwing 
salt over his shoulder. 

A member of the Roman Catholic Church, unless granted a 
dispensation, is required by his religion to abstain from eating 
meat on Fridays and during Lent. If he fails to observe the rule he 
sins, and must proceed, as in any other sin, to confess and obtain 
absolution. Different as this is in important ways from the rule 
about spilling salt, it can and must for scientific purposes be 
regarded as belonging to the same general class. Eating meat on 
Friday produces in the person who does so an undesirable change 
of ritual status which requires to be remedied by fixed appropriate 

We may add to these examples two others from other societies. 
If you turn to the fifth chapter of Leviticus you will find that 
amongst the Hebrews if a 'soul' touch the carcase of an unclean 
beast or of unclean cattle, or of unclean creeping things, even if 
he is unaware that he does so, then he is unclean and guilty and 
has sinned. When he becomes aware of his sin he must confess 
that he has sinned and must take a trespass offering — a female 
from the flock, a lamb or a kid of the goats — which the priest 


shall sacrifice to make an atonement for the sin so that it shall be 
forgiven him. Here the change in ritual status through touching 
an unclean carcase is described by the terms 'sin', 'unclean' 
and 'guilty'. 

In the Kikuyu tribe of East Africa the word thahu denotes 
the undesirable ritual status that results from failure to observe 
rules of ritual avoidance. It is believed that a person who is thahu 
will be ill and will probably die unless he removes the thahu 
by the appropriate ritual remedies, which in all serious cases 
require the services of a priest or medicine man. Actions which 
produce this condition are touching or carrying a corpse, stepping 
over a corpse, eating food from a cracked pot, coming in contact 
with a woman's menstrual discharge, and many others. Just 
as among the Hebrews a soul may unwittingly be guilty of sin 
by touching in ignorance the carcase of an unclean animal, so 
amongst the Kikuyu a man may become thahu without any 
voluntary act on his part. If an elder or a woman when coming 
out of the hut slips and falls down on the ground, he or she is 
thahu and lies there until some of the elders of the neighbourhood 
come and sacrifice a sheep. If the side-pole of a bedstead breaks, 
the person lying on it is thahu and must be purified. If the drop- 
pings of a kite or crow fall on a person he is thahu, and if a hyaena 
defaecates in a village, or a jackal barks therein, the village and 
its inhabitants are thahu. 

I have purposely chosen from our society two examples of 
ritual avoidances which are of very difi^erent kinds. The rule 
against eating meat on Friday or in Lent is a rule of religion, as is 
the rule, where it is recognised, against playing golf or tennis on 
Sunday. The rule against spilling salt, I suppose it will be agreed, 
is non- religious. Our language permits us to make this distinction 
very clearly, for infractions of the rules of religion are sins, while 
the non-religious avoidances are concerned v.ith good and bad 
luck. Since this distinction is so obvious to us it might be thought 
that we should find it in other societies. My own experience is 
that in some of the societies with which I am acquainted this 
distinction between sinful acts and acts that bring bad luck 
cannot be made. Several anthropologists, however, have attempted 
to classify rites into two classes, religious rites and magical rites. 

For Emile Durkheim the essential distinction is that religious 
rites are obligatory within a religious society or church, while 

TABOO 137 

magical rites are optional. A person who fails in religious obser- 
vances is guilty of wrong-doing, whereas one who does not observe 
the precautions of magic or those relating to luck is simply acting 
foolishly. This distinction is of considerable theoretical im- 
portance. It is difficult to apply in the study of the rites of 
simple societies. 

Sir James Frazer defines religion as 'a propitiation or con- 
ciliation of superhuman powers which are believed to control 
nature and man', and regards magic as the erroneous application 
of the notion of causality. If we apply this to ritual prohibitions, 
we may regard as belonging to religion those rules the infraction 
of which produces a change of ritual status in the individual 
by offending the superhuman powers, whereas the infraction 
of a rule of magic would be regarded as resulting immediately 
in a change of ritual status, or in the misfortune that follows, by a 
process of hidden causation. Spilling salt, by Sir James Frazer's 
definition, is a question of magic, while eating meat on Friday is a 
question of religion. 

An attempt to apply this distinction systematically meets with 
certain difficulties. Thus with regard to the Maori Sir James 
Frazer states that 'the ultimate sanction of the taboo, in other 
words, that which engaged the people to observe its command- 
ments, was a firm persuasion that any breach of those command- 
ments would surely and speedily be punished by an atua or ghost, 
who would afflict the sinner with a painful malady till he died'. 
This would seem to make the Polynesian taboo a matter of religion, 
not of magic. But my own observation of the Polynesians suggests 
to me that in general the native conceives of the change in his 
ritual status as taking place as the immediate result of such an act 
as touching a corpse, and that it is only when he proceeds to 
rationalise the whole system of taboos that he thinks of the gods 
and spirits — the atiia — ^as being concerned. Incidentally it should 
not be assumed that the Polynesian word atua or otiia always 
refers to a personal spiritual being. 

Of the various ways of distinguishing magic and religion I 
will mention only one more. For Professor Malinowski a rite is 
magical when 'it has a definite practical purpose which is known 
to all who practise it and can be easily elicited from any native 
informant', while a rite is religious if it is simply expressive and 
has no purpose, being not a means to an end but an end in itself. 


A difficulty in applying this criterion is due to uncertainty as t 
what is meant by 'definite practical purpose'. To avoid the ba 
luck which results from spilling salt is, I suppose, a practica; 
purpose though not very definite. The desire to please God ii 
all our actions and thus escape some period of Purgatory i 
perhaps definite enough, but Professor Malinowski may regard i 
as not practical. What shall we say of the desire of the Polynesia 
to avoid sickness and possible death which he gives as his reaso 
for not touching chiefs, corpses and newly-born babies? 

Seeing that there is this absence of agreement as to th 
definitions of magic and religion and the nature of the distinctic; 
between them, and seeing that in many instances whether v 
call a particular rite magical or religious depends on which ' 
the various proposed definitions we accept, the only soue 
procedure, at any rate in the present state of anthropologic 
knowledge, is to avoid as far as possible the use of the terms ; 
question until there is some general agreement about thei 
Certainly the distinctions made by Durkheim and Frazer ai 
Malinowski may be theoretically significant, even though th^ 
are difficult to apply universally. Certainly, also, there is ne< 
for a systematic classification of rites, but a satisfacto? 
classification will be fairly complex and a simple dichoton 
between magic and religion does not carry us very far towards : 

Another distinction which we make in our own sociei 
within the field of ritual avoidances is between the holy and tl 
unclean. Certain things must be treated with respect because the 
are holy, others because they are unclean. But, as Robertsc 
Smith and Sir James Frazer have shown, there are many societio 
in which this distinction is entirely unrecognised. The Polynesian 
for example, does not think of a chief or a temple as holy and 
corpse as unclean. He thinks of them all as things dangerous. A 
example from Hawaii will illustrate this fundamental identit 
of holiness and uncleanness. There, in former times, if a commom 
committed incest with his sister he became kapu (the Hawaiian; 
form of tabu). His presence was dangerous in the extreme for tbj. 
whole community, and since he could not be purified he was pi 
to death. But if a chief of high rank, who, by reason of his rank wa^ 
of course, sacred (kapu), married his sister he became still more s(V 
An extreme sanctity or untouchability attached to a chief born of I 
brother and sister who were themselves the children of a brothel 

TABOO 139 

and sister. The sanctity of such a chief and the uncleanness of the 
person put to death for incest have the same source and are 
the same thing. They are both denoted by saying that the person 
is kapu. In studying the simpler societies it is essential that we 
should carefully avoid thinking of their behaviour and ideas in 
terms of our own ideas of holiness and uncleanness. Since most 
people find this difficult it is desirable to have terms which we can 
use that do not convey this connotation. Durkheim and others have 
used the word 'sacred' as an inclusive term for the holy and the 
unclean together. This is easier to do in Frtnch than in English, 
and has some justification in the fact that the Latin sacer did apply 
to holy things such as the gods and also to accursed things such as 
persons guilty of certain crimes. But there is certainly a tendency 
in English to identify sacred with holy. I think that it will greatly 
aid clear thinking if we adopt some wide inclusive term which does 
not have any undesirable connotation. I venture to propose the 
term 'ritual value'. 

Anything — a person, a material thing, a place, a word or name, 
an occasion or event, a day of the week or a period of the year — 
which is the object of a ritual avoidance or taboo can be said to have 
ritual value. Thus in Polynesia chiefs, corpses and ncMly-born 
babies have ritual value. For some people in England salt has 
ritual value. For Christians all Sundays and Good Friday have 
ritual value, and for Jews all Saturdays and the Day of Atonement. 
The ritual value is exhibited in the behaviour adopted towards 
the object or occasion in question. Ritual values are exhibited 
not only in negative ritual but also in positive ritual, being 
possessed by the objects towards which positive rites are directed 
and also by objects, words or places used in the rites. A large class 
of positive rites, those of consecration or sacralisation, have for 
their purpose to endow objects with ritual value. It may be noted 
that in general anything that has value in positive ritual is also the 
object of some sort of ritual avoidance or at the very least of ritual 

The word 'value', as I am using it, always refers to a relation 
between a subject and an object. The relation can be stated in two 
ways by saying either that the object has a value for the subject, 
or that the subject has an interest in the object. We can use the 
terms in this way to refer to any act of behaviour towards an object. 
The relation is exhibited in and defined by the behaviour. The 


words 'interest' and 'value' provide a convenient shorthand by 
which we can describe the reaUty, which consists of acts of be- 
haviour and the actual relations between subjects and objects 
which those acts of behaviour reveal. If Jack loves Jill, then Jill 
has the value of a loved object for Jack, and Jack has a recognisable 
interest in Jill. When I am hungry I have an interest in food, and a 
good meal has an immediate value for me that it does not have at 
other times. My toothache has a value to me as something that I 
am interested in getting rid of as quickly as possible, 

A social system can be conceived and studied as a system of 
values. A society consists of a number of individuals bound together 
in a network of social relations. A social relation exists between 
two or more persons when there is some harmonisation of their 
individual interests, by some convergence of interest and by 
limitation or adjustment of divergent interests. An interest is 
always the interest of an individual. Two individuals may have 
similar interests. Similar interests do not in themselves constitute 
a social relation; two dogs may have a similar interest in the same 
bone and the result may be a dog-fight. But a society cannot exist 
except on the basis of a certain measure of similarity in the 
interests of its members. Putting this in terms of value, the 
first necessary condition of the existence of a society is that 
the individual members shall agree in some measure in the values 
that they recognise. 

Any particular society is characterised by a certain set of 
values — moral, aesthetic, economic, etc. In a simple society there 
is a fair amount of agreement amongst the members in their 
evaluations, though of course the agreement is never absolute. 
In a complex modern society we find much more disagreement 
if we consider the society as a whole, but we may find a closer 
measure of agreement amongst the members of a group or class 
within the society. 

While some measure of agreement about values, some simi- 
larity of interests, is a prerequisite of a social system, social 
relations involve more than this. They require the existence of 
common interests and of social values. When two or more persons 
have a common interest in the same object and are aware of their 
community of interest a social relation is established. They form, 
whether for a moment or for a long period, an association, and the 
object may be said to have a social value. For a man and his wife 

TABOO 141 

the birth of a child, the child itself and its well-being and happi- 
ness or its death, are objects of a common interest which binds 
them together and they thus have, for the association formed by 
the two persons, social value. By this definition an object can only 
have a social value for an association of persons. In the simplest 
possible instance we have a triadic relation; Subject i and Subject 2 
are both interested in the same way in the Object and each of the 
Subjects has an interest in the other, or at any rate in certain 
items of the behaviour of the other, namely those directed towards 
the object. To avoid cumbersome circumlocutions it is convenient 
to speak of the object as having a social value for any one subject 
involved in such a relation, but it must be remembered that this 
is a loose way of speaking. 

It is perhaps necessary for the avoidance of misunderstanding 
to add that a social system also requires that persons should be 
objects of interest to other persons. In relations of friendship 
or love each of two persons has a value for the other. In certain 
kinds of groups each member is an object of interest for all the 
others, and each member therefore has a social value for the group 
as a whole. Further, since there are negative values as well as 
positive, persons may be united or associated by their antagonism 
to other persons. For the members of an anti-Comintern pact the 
Comintern has a specific social value. 

Amongst the members of a society we find a certain measure of 
agreement as to the ritual value they attribute to objects of different 
kinds. We also find that most of these ritual values are social 
values as defined above. Thus for a local totemic clan in Australia 
the totem-centres, the natural species associated with them, i.e. 
the totems, and the myths and rites that relate thereto, have a 
specific social value for the clan; the common interest in them 
binds the individuals together into a firm and lasting association. 

Ritual values exist in every known society, and show an im- 
mense diversity as we pass from one society to another. The 
problem of a natural science of society (and it is as such that I 
regard social anthropology) is to discover the deeper, not im- 
mediately perceptible, uniformities beneath the superficial differ- 
ences. This is, of course, a highly complex problem which will 
require the studies begun by Sir James Frazer and others to be 
continued by many investigators over many years. The ultimate 
aim should be, I think, to find some relatively adequate answer 


to the question — What is the relation of ritual and ritual values 
to the essential constitution of human society? I have chosen a 
particular approach to this study which I beheve to be promising — 
to investigate in a few societies studied as thoroughly as possible 
the relations of ritual values to other values including moral and 
aesthetic values. In the present lecture, however, it is only one 
small part of this study in which I seek to interest you — the 
question of a relation between ritual values and social values. 

One way of approaching the study of ritual is by the con- 
sideration of the purposes or reasons for the rites. If one examines 
the literature of anthropology one finds this approach very 
frequently adopted. It is by far the least profitable, though the 
one that appeals most to common sense. Sometimes the purpose 
of a rite is obvious, or a reason may be volunteered by those who 
practise it. Sometimes the anthropologist has to ask the reason, 
and in such circumstances it may happen that different reasons 
are given by different informants. What is fundamentally the 
same rite in two different societies may have different purposes 
or reasons in the one and in the other. The reasons given by 
the members of a community for any custom they observe are 
important data for the anthropologist. But it is to fall into grievous 
error to suppose that they give a valid explanation of the custom. 
What is entirely inexcusable is for the anthropologist, when he 
cannot get from the people themselves a reason for their behaviour 
which seems to him satisfactory, to attribute to them some purpose 
or reason on the basis of his own preconceptions about human 
motives. I could adduce many instances of this from the literature 
of ethnography, but I prefer to illustrate what I mean by an 

A Queenslander met a Chinese who was taking a bowl of 
cooked rice to place on his brother's grave. The Australian in 
jocular tones asked if he supposed that his brother would come 
and eat the rice. The reply was 'No! We offer rice to people as an 
expression of friendship and affection. But since you speak as you 
do I suppose that you in this country place flowers on the graves 
of your dead in the belief that they will enjoy looking at them and 
smelling their sweet perfume.' 

So far as ritual avoidances are concerned the reasons for them 
may vary from a very vague idea that some sort of misfortune or 
ill-luck, not defined as to its kind, is likely to befall anyone who isit 

TABOO 143 

fails to observe the taboo, to a belief that non-observance will 
produce some quite specific and undesirable result. Thus an 
Australian aborigine told me that if he spoke to any woman who 
stood in the relation of mother-in-law to him his hair would turn 

The very common tendency to look for the explanation of 
ritual actions in their purpose is the result of a false assimilation 
of them to what may be called technical acts. In any technical 
activity an adequate statement of the purpose of any particular 
act or series of acts constitutes by itself a sufficient explanation. 
But ritual acts differ from technical acts in having in all instances 
some expressive or symbolic element in them. 

A second approach to the study of ritual is therefore by a 
consideration not of their purpose or reason but of their meaning. 
I am here using the words symbol and meaning as coincident. 
Whatever has a meaning is a symbol and the meaning is whatever 
is expressed by the symbol. 

But how are we to discover meanings? They do not lie on the 
surface. There is a sense in which people always know the meaning 
of their own symbols, but they do so intuitively and can rarely 
express their understanding in words. Shall we therefore be 
reduced to guessing at meanings as some anthropologists have 
guessed at reasons and purposes? I think not. For as long as we 
admit guess-work of any kind social anthropology^ cannot be a 
science. There are, I believe, methods of determining, with some 
fair degree of probability, the meanings of rites and other s}Tnbols. 

There is still a third approach to the study of rites. We can 
consider the effects of the rite — not the effects that it is supposed 
to produce by the people who practise it but the effects that it does 
actually produce. A rite has immediate or direct effects on the 
persons who are in any way directly concerned in it, which we may 
call, for lack of a better term, the psychological effects. But there 

1 In case it may be thought that this is an inadequate supernatural punish- 
ment for a serious breach of rules of proper behaviour a few words of ex- 
planation are necessary. Grey hair comes with old age and is thought to be 
usually associated with loss of sexual potency. It is thus premature old age 
with its disadvantages but without the advantages that usually accompany 
seniority that threatens the man who fails to observe the rules of avoidance, 
♦be other hand when a man's hair is grey and his wife's mother has passed 
perge of child-bearing the taboo is relaxed so that the relatives may talk 
i Sacher if they wish. 


are also secondary effects upon the social structure, i.e. the network 
of social relations binding individuals together in an ordered life. 
These we may call the social effects. By considering the psychological 
effects of a rite we may succeed in defining its psychological function; 
by considering the social effects we may discover its social function. 
Clearly it is impossible to discover the social function of a rite with- 
out taking into account its usual or average psychological effects. 
But it is possible to discuss the psychological effects while more 
or less completely ignoring the more remote sociological effects, 
and this is often done in what is called 'functional anthropology'. 

Let us suppose that we wish to investigate in Australian tribes 
the totemic rites of a kind widely distributed over a large part of 
the continent. The ostensible purpose of these rites, as stated 
by the natives themselves, is to renew or maintain some part of 
nature, such as a species of animal or plant, or rain, or hot or 
cold weather. With reference to this purpose we have to say that 
from our point of view the natives are mistaken, that the rites 
do not actually do what they are believed to do. The rain-making 
ceremony does not, we think, actually bring rain. In so far as the 
rites are performed for a purpose they are futile, based on erroneous 
belief. I do not believe that there is any scientific value in attempts 
to conjecture processes of reasoning which might be supposed to 
have led to these errors. 

The rites are easily perceived to be symbolic, and we may 
therefore investigate their meaning. To do this we have to examine 
a considerable number of them and we then discover that there is 
a certain body of ritual idiom extending from the west coast of the 
continent to the east coast with some local variations. Since each 
rite has a myth associated with it we have similarly to investigate 
the meanings of the myths. As a result we find that the meaning of 
any single rite becomes clear in the light of a cosmolog}^ a body 
of ideas and beliefs about nature and human society, which, so 
far as its most general features are concerned, is current in all 
Australian tribes. 

The immediate psychological effects of the rites can be to 
some extent observed by watching and talking to the performers. 
The ostensible purpose of the rite is certainly present in their 
minds, but so also is that complex set of cosmological beliefs by 
reference to which the rite has a meaning. Certainly a pc oon 
performing the rite, even if, as sometimes happens, he perfon-lis it 

TABOO 145 

alone, derives therefrom a definite feeling of satisfaction, but it 
would be entirely false to imagine that this is simply because he 
believes that he has helped to provide a more abundant supply 
of food for himself and his fellow-tribesmen. His satisfaction is in 
having performed a ritual duty, we might say a religious duty. 
Putting in my own words what I judge, from my own observations, 
to express what the native feels, I would say that in the performance 
of the rite he has made that small contribution, which it is both 
his privilege and his duty to do, to the maintenance of that order 
of the universe of which man and nature are interdependent 
parts. The satisfaction which he thus receives gives the rite a 
special value for him. In some instances with which I am ac- 
quainted of the last survivor of a totemic group who still continues 
to perform the totemic rites by himself, it is this satisfaction that 
constitutes apparently the sole motive for his action. 

To discover the social function of the totemic rites we have 
to consider the whole body of cosmological ideas of which each 
rite is a partial expression. I believe that it is possible to show 
that the social structure of an Australian tribe is connected in a 
very special way with these cosmological ideas and that the main- 
tenance of its continuity depends on keeping them alive, by 
their regular expression in myth and rite. 

Thus any satisfactory study of the totemic rites of Australia 
must be based not simply on the consideration of their ostensible 
purpose and their psychological function, or on an analysis of the 
motives of the individuals who perform the rites, but on the 
discovery of their meaning and of their social function. 

It may be that some rites have no social function. This may 
be the case with such taboos as that against spilling salt in our own 
society. Nevertheless, the method of investigating rites and ritual 
values that I have found most profitable during work extending 
over more than thirty years is to study rites as symbolic expressions 
and to seek to discover their social functions. This method is not 
new except in so far as it is applied to the comparative study of 
many societies of diverse types. It was applied by Chinese thinkers 
to their own ritual more than twenty centuries ago. 

In China, in the fifth and sixth centuries B.C., Confucius and 
his followers insisted on the great importance of the proper 
performance of ritual, such as funeral and mourning rites and 
sacrifices. After Confucius there came the reformer Mo Ti who 



taught a combination of altruism — love for all men — and utili- 
tarianism. He held that funeral and mourning rites were useless 
and interfered with useful activities and should therefore be 
abolished or reduced to a minimum. In the third and second 
centuries B.C., the Confucians, Hsiin Tze and the compilers of the 
Li Chi (Book of Rites), replied to Mo Ti to the eifect that though 
these rites might have no utilitarian purpose they none the less 
had a very important social function. Briefly the theory is that the 
rites are the orderly (the Li Chi says the beautified) expression of 
feelings appropriate to a social situation. They thus serve to 
regulate and refine human emotions. We may say that partaking 
in the performance of rites serves to cultivate in the individual 
sentiments on whose existence the social order itself depends. 
~" -. Let us consider the meaning and social function of an extremely 
simple example of ritual. In the Andaman Islands when a woman 
is expecting a baby a name is given to it while it is still in the 
I womb. From that time until some weeks after the baby is born 
1 nobody is allowed to use the personal name of either the father or 
{ the mother; they can be referred to by teknonymy, i.e. in terms of 
\ their relation to the child. During this period both the parents are 
' required to abstain from eating certain foods which they may 
i^^^ely eat at other times. 

/""^ did not obtain from the Andamanese any statement of the 
(purpose or reason for this avoidance of names. Assuming that the 
lact is symbolic, what method, other than that of guessing, is there 
jof arriving at the meaning? I suggest that we may start with a 
general working hypothesis that when, in a single society, the 
(Same s)rmbol is used in different contexts or on different kinds of 
[occasions there is some common element of meaning, and that 
by comparing together the various uses of the symbol we may be 
! able to discover what the common element is. This is precisely the 
j method that we adopt in studying an unrecorded spoken language 
/ in order to discover the meanings of words and morphemes. 
' In the Andamans the name of a dead person is avoided from 
the occurrence of the death to the conclusion of mourning; the 
name of a person mourning for a dead relative is not used; there is 
avoidance of the name of a youth or girl who is passing through 
the ceremonies that take place at adolescence; a bride or bride- 
groom is not spoken of or to by his or her own name for a short 
time after the marriage. For the Andamanese the personal name 

TABOO 14* 

is a symbol of the social personality, i.e. of the position that an 
individual occupies in the social structure and the social life. 
The avoidance of a personal name is a symbolic recognition of the 
fact that at the time the person is not occupying a normal position 
in the social life. It may be added that a person whose name is 
thus temporarily out of use is regarded as having for the time an 
abnormal ritual status. 

Turning now to the rule as to avoiding certain foods, if the 
Andaman Islanders are asked what would happen if the father 
or mother broke his taboo the usual answer is that he or she would 
be ill, though one or two of my informants thought it might per- 
haps also affect the child. This is simply one instance of a standard 
formula which applies to a number of ritual prohibitions. Thus 
persons in mourning for a relative may not eat pork and turtle, the 
most important flesh foods, and the reason given is that if they 
did they would be ill. 

To discover the meaning of this avoidance of foods by the 
parents we can apply the same method as in reference to the 
avoidance of their names. There are similar rules for mourners, 
for women during menstruation, and for youths and girls during 
the period of adolescence. But for a full demonstration we have 
to consider the place of foods in Andamanese ritual as a w'hole, 
and for an examination of this I must refer to v. hat I have already 
written on the subject. 

I should like to draw your attention to another point in the 
method by W'hich it is possible to test our hypotheses as to the 
meanings of rites. We take the different occasions on which two 
rites are associated together, for example the association of the 
avoidance of a person's name with the avoidance by that person of 
certain foods, which we find in the instance of mourners on the one 
hand and the expectant mother and father on the other. We must 
assume that for the Andamanese there is some important similarity 
between these two kinds of occasions — birth and death — by 
virtue of which they have similar ritual values. We cannot rest 
content with any interpretation of the taboos at childbirth unless 
there is a parallel interpretation of those relating to mourners. 
In the terms I am using here we can say that in the Andamans the 
relatives of a recently dead person, and the father and mother of 
a child that is about to be, or has recently been, born, are in an 
abnormal ritual status. This is recognised or indicated by the 


avoidance of their names. They are regarded as hkely to suffer 
some misfortune, some bad luck, if you will, unless they observe 
certain prescribed ritual precautions of which the avoidance of 
certain foods is one. In the Andaman Islands the danger in such 
instances is thought of as the danger of illness. This is the case also 
with the Polynesian belief about the ritual status of anyone who 
has touched a corpse or a newly-born baby. It is to be noted that 
for the Polynesians as well as for the Andamanese the occasion of 
a birth has a similar ritual value to that of a death. 

The interpretation of the taboos at childbirth at which we 
arrive by studying it in relation to the whole system of ritual 
values of the Andamanese is too complex to be stated here in 
full. Clearly, however, they express, in accordance with Anda- 
manese ritual idiom, a common concern in the event. The parents 
show their concern by avoiding certain foods; their friends show 
theirs by avoiding the parents' personal names. By virtue of 
these taboos the occasion acquires a certain social value, as that 
term has been defined above. 

There is one theory that might seem to be applicable to our 
example. It is based on a~-hy;pothesis as to the psychological 
function of a class of rites. The theory is that in certain circum- 
stances the individual humaiTbeing is anxious about the outcome 
of some event or activity because it depends to some extent on 
conditions that he cannot control by any technical means. He 
therefore observes some rite which, since he believes it will ensure 
good luck, serves to reassure hinDThus an aeronaut takes with 
him in a plane a mascot which he believes will protect him from 
accident and thus carries out his flight with confidence. 

The theory has a respectable antiquity. It was perhaps implied 
in the Primus in orbe deos fecit timor of Petronius and Statins. 
It has taken various forms from Hume's explanation of religion 
to Malinowski's explanation of Trobriand magic. It can be made 
so plausible by a suitable selection of illustrations that it is neces- 
sary to examine it with particular care and treat it with reasonable 
scepticism. For there is always the danger that we may be taken 
in by the plausibility of a theory that ultimately proves to be 

I think that for certain rites it would be easy to maintain with 
equal plausibility an exactly contrary theory, namely, that if it 
were not for the existence of the rite and the beliefs associated 

TABOO 149 

with it the individual would feel no anxiety, and that the psycho- 
logical effect of the rite is to create in him a sense of insecurity or 
danger. It seems very unlikely that an Andaman Islander would 
think that it is dangerous to eat dugong or pork or turtle meat if it 
were not for the existence of a specific body of ritual the ostensible 
purpose of which is to protect him from those dangers. Many 
hundreds of similar instances could be mentioned from all over 
the world. 

Thus, while one anthropological theory is that magic and 
religion give men confidence, comfort and a sense of security,^ 
it could equally well be argued that they give men fears and anxieties 
from which they would otherwise be free — the fear of black 
magic or of spirits, fear of God, of the Devil, of Hell. 

Actually in our fears or anxieties as well as in our hopes we 
are conditioned (as the phrase goes) by the community in which 
we live. And it is largely by the sharing of hopes and fears, by 
what I have called common concern in events or eventualities, that 
human beings are linked together in temporary or permanent 

To return to the Andamanese taboos at childbirth, there are 
difficulties in supposing that they are means by which parents 
reassure themselves against the accidents that may interfere with a 
successful delivery. If the prospective father fails to observe the 
i food taboo it is he who will be sick, according to the general 
Andamanese opinion. Moreover, he must continue to observe 
the taboos after the child is safely delivered. Further, how are we 
to provide a parallel explanation of the similar taboos observed 
by a person mourning for a dead relative? 

The taboos associated with pregnancy and parturition are 
often explained in terms of the hypothesis I have mentioned. 
A father, naturally anxious at the outcome of an event over which 
he does not have a technical control and which is subject to 
hazard, reassures himself by observing some taboo or carr^^ing out 
some magical action. He may avoid certain foods. He may avoid 
making nets or tying knots, or he may go round the house untying 
all knots and opening any locked or closed boxes or containers. 

I wish to arouse in your minds, if it is not already there, a 
suspicion that both the general theory and this special application 

^ This theory has been formulated by Loisy, and for magic has been adopted 
by MaHnowski. 


of it do not give the whole truth and indeed may not be true at all. 
Scepticism of plausible but unproved hypotheses is essential in 
every science. There is at least good ground for suspicion in the 
fact that the theory has so far been considered in reference to 
facts that seem to fit it, and no systematic attempt has been made, 
so far as I am aware, to look for facts that do not fit. That there 
are many such I am satisfied from my own studies. 

/The alternative hypothesis which I am presenting for con- 
sideration is as follows. In a given community it is appropriate 
that an expectant father should feel concern or at least should 
make an appearance of doing so. Some suitable symbolic ex- 
pression of his concern is found in terms of the general ritual 
or symbolic idiom of the society, and it is felt generally that a man 
in that situation ought to carry out the symbolic or ritual actions 
or abstentions. For every rule that ought to be observed there must 
be some sort of sanction or reason. For acts that patently affect 
other persons the moral and legal sanctions provide a generally 
sufficient controlling force upon the individual. For ritual ob- 
ligations conformity and rationalisation are provided by the ritual 
sanctions. The simplest form of ritual sanction is an accepted 
belief that if rules of ritual are not observed some undefined 
misfortune is likely to occur. In many societies the expected 
danger is somewhat more definitely conceived as a danger of 
sickness or, in extreme cases, death. In the more specialised forms 
of ritual sanction the good results to be hoped for or the bad 
results to be feared are more spectfically defined in reference to 
the occasion or meaning of the ritual./ 

The theory is not concerned with the historical origin of 
ritual, nor is it another attempt to explain ritual in terms of human 
psychology; it is a hypothesis as to the relation of ritual and 
ritual values to the essential constitution of human society, i.e. 
to those invariant general characters which belong to all human 
societies, past, present and future. It rests on the recognition of 
the fact that while in animal societies social coaptation depends on 
instinct,Jn^human societies it depends upon the efficacy of symbols 
of many different kinds. The theory I am advancing must there- 
fore, for a just estimation of its value, be considered in its place 
in a general theory of symbols and their social efficacy. 

By this theory the Andamanese taboos relating to childbirth 
are the obligatory recognition in a standardised symbolic form 

TABOO 151 

of the significance and importance of the event to the parents and 
to the community at large. They thus serve to fix the social value 
of occasions of this kind. Similarly I have argued in another place 
that the Andamanese taboos relating to the animals and plants 
used for food are means of affixing a definite social value to food, 
based on its social importance. The social importance of food is 
not that it satisfies hunger, but that in such a community as an 
Andamanese camp or village an enormously large proportion 
of the activities are concerned with the getting and consuming of 
food, and that in these activities, with their daily instances of 
collaboration and mutual aid, there continuously occur those 
inter-relations of interests which bind the individual men, women 
and children into a society. 

I believe that this theory can be generalised and with suitable 
modifications wuU be found to apply to a vast number of the taboos 
of different societies. My theory would go further for I would hold, 
as a reasonable working hypothesis, that we have here the primary 
basis of all ritual and therefore of religion and magic, however 
those may be distinguished. The primary basis of ritual, so the 
formulation would run, is the attribution of ritual value to objects 
and occasions which are either themselves objects of important 
common interests linking together the persons of a community 
or are symbolically representative of such objects. To illustrate 
what is meant by the last part of this statement two illustrations 
may be off^ered. In the Andamans ritual value is attributed to the 
cicada, not because it has any social importance itself but because 
it symbolically represents the seasons of the year which do have 
importance. In some tribes of Eastern Australia the god Baiame 
is the personification, i.e. the symbolical representative, of the 
moral law of the tribe, and the rainbow-serpent (the Australian 
equivalent of the Chinese dragon) is a symbol representing 
growth and fertility in nature. Baiame and the rainbow-serpent 
in their turn are represented by the figures of earth which are 
made on the sacred ceremonial ground of the initiation cere- 
monies and at which rites are performed. The reverence that the 
Australian shows to the image of Baiame or towards his name is 
the symbolic method of fixing the social value of the moral law, 
particularly the laws relating to marriage. 

In conclusion let me return once more to the work of the 
anthropologist whom we are here to honour. Sir James Frazer, in 


his Psyche's Task and in his other works, set himself to show how, 
in his own words, taboos have contributed to build up the complex 
fabric of society. He thus initiated that functional study of ritual 
to which I have in this lecture and elsewhere attempted to make 
some contribution. But there has been a shift of emphasis. Sir 
James accounted for the taboos of savage tribes as the application 
in practice of beliefs arrived at by erroneous processes of reasoning, 
and he seems to have thought of the effects of these beliefs in 
creating or maintaining a stable orderly society as being accidental. 
My own view is that the negative and positive rites of savages 
exist and persist because they are part of the mechanism by which 
an orderly society maintains itself in existence, serving as they do 
to establish certain fundamental social values. The beliefs by 
which the rites themselves are justified and given some sort of 
consistency are the rationalisations of symbolic actions and of the 
sentiments associated with them. I would suggest that what Sir 
James Frazer seems to regard as the accidental results of magical 
and religious beliefs really constitute their essential function and 
the ultimate reason for their existence. 


The theory of ritual outlined in this lecture was first worked 
out in 1908 in a thesis on the Andaman Islanders. It was written 
out again in a revised and extended form in 191 3 and appeared 
in print in 1922. Unfortunately the exposition contained in 
The Andaman Islanders is evidently not clear, since some of my 
critics have failed to understand what the theory is. For example, 
it has been assumed that by 'social value' I mean 'utility'. 

The best treatment of the subject of value with which I am 
acquainted is Ralph Barton Perry's General Theory of Value, 1926. 
For the Chinese theory of ritual the most easily accessible account 
is in chapter xiv of Fung Yu-lan's History of Chinese Philosophy, 
1937. The third chapter, on the uses of symboHsm, of Whitehead's 
Symbolism, its Meaning and Effect, is an admirable brief introduc- 
tion to the sociological theory of symbolism. 

One very important point that could not be dealt with in the 
lecture is that indicated by Whitehead in the following sentence — 
'No account of the uses of symbolism is complete without the 
recognition that the symbolic elements in life have a tendency to 
run wild, like the vegetation in a tropical forest.' 



K 1 IHE Royal Anthropological Institute has honoured me 
I with an invitation to deliver the Henry Myers Lecture 
JL. on the role of religion in the development of human 
society. That is an important and complex subject, about which 
it is not possible to say very much in a single lecture, but as it is 
hoped that this may be only the first of a continuing series of 
lectures, in which different lecturers will each offer some con- 
tribution, I think that the most useful thing I can do is to indicate 
certain lines along which I believe that an enquir}^ into this 
problem can be profitably pursued. 

The usual way of looking at religions is to regard all of them, 
or all except one, as bodies of erroneous beliefs and illusory 
practices. There is no doubt that the histor}' of religions has been 
in great part a history of error and illusion. In all ages men have 
hoped that by the proper performance of religious actions or 
observances they would obtain some specific benefit: health and 
long life, children to carry on their line, material well-being, 
success in hunting, rain, the growth of crops and the multiplica- 
tion of cattle, victory in war, admission of their souls after death 
to a paradise, or inversely, release by the extinction of personality 
from the round of reincarnation. We do not believe that the rain- 
making rites of savage tribes really produce rain. Nor do we 
believe that the initiates of the ancient mysteries did actually 
attain through their initiation an immortality denied to other men. 
When we regard the religions of other peoples, or at least those 
of what are called primitive peoples, as systems of erroneous and 
illusor}'^ beliefs, we are confronted with the problem of ho\^- these 
beliefs came to be formulated and accepted. It is to this problem 
that anthropologists have given most attention. My personal 
opinion is that this method of approach, even though it may seem 
the most direct, is not the one most likely to lead to a real under- 
standing of the nature of religions. 

^ The Henry Myers Lecture, 1945. 


There is another way in which we may approach the study of 
religions. We may entertain as at least a possibility the theory that 
any religion is an important or even essential part of the social 
machinery, as are morality and law, part of the complex system 
by which human beings are enabled to live together in an orderly 
arrangement of social relations. From this point of view we deal 
not with the origins but with the social functions of religions, 
i.e. the contribution that they make to the formation and main- 
tenance of a social order. There are many persons who would say 
that it is only true religion (i.e. one's own) that can provide the 
foundation of an orderly social life. The hypothesis we are con- 
sidering is that the social function of a religion is independent 
of its truth or falsity, that religions which we think to be erroneous 
or even absurd and repulsive, such as those of some savage tribes, 
may be important and effective parts of the social machinery, 
and that without these 'false' religions social evolution and 
the development of modern civilisation would have been 

The hypothesis, therefore, is that in what we regard as false 
religions, ihough,, the performance of religious rites does not 
actually produce the effects that are expected or hoped for b_y those 
who perform or take part in them, they have other effects, some 
at least of which may be socially valuable. 

How are we to set to work to test this hypothesis? It is of no 
use thinking in terms of religion in general, in the abstract, and 
society in the abstract. Nor is it adequate to consider some one 
religion, particularly if it is the one in which we have been brought 
up and about which we are likely to be prejudiced one way or 
another. The only method is the experimental method of social 
anthropology, and that means that we must study in the light 
of our hypothesis a sufficient number of diverse particular 
religions or religious cults in their relation to the particular 
societies in which they are found. This is a task not for one person i 
but for a number. 1 

Anthropologists and others have discussed at length the ; 
question of the proper definition of religion. I do not intend to 
deal with that controversial subject on this occasion. But there 
are some points that must be considered. I shall assume that any 
religion or any religious cult normally involves certain ideas or 
beliefs on the one hand, and on the other certain observances 


These observances, positive and negative, i.e. actions and absten- 
tions, I shall speak of as rites. 

In European countries, and more particularly since the 
Reformation, religion has come to be considered as primarily a 
matter of belief. This is itself a phenomenon which needs to be 
explained, I think, in terms of social development. We are con- 
cerned here only with its effects on the thinking of anthropologists. 
Among many of them there is a tendency to treat belief as primary: 
rites are considered as the results of beliefs. They therefore con- 
centrate their attention on tr^^ing to explain the beliefs by hypo- 
theses as to how they may have been formed and adopted. 

To my mind this is the product of false psychology. For 
example, it is sometimes held that funeral and mourning rites 
are the result of a belief in a soul surviving death. If we must 
talk in terms of cause and effect, I would rather hold the view that 
the belief in a surviving soul is not the cause but the effect of the 
rites. Actually the cause-effect analysis is misleading. What really 
happens is that the rites and the justifying or rationalising beliefs 
develop together as parts, of a coherent whole^ But in this develop- 
mentit_is action or the need of action that controls or determines 
belief rather than the other way about. The actions themselves 
are symbolic expressions of sentiments. 

My suggestion is that in attempting to understand a religion 
it is on the rites rather than on the beliefs that we should first 
concentrate our attention. Much the same view is taken by Loisy, 
who justifies his selection of sacrificial rites as the subject of his 
analysis of religion by saying that rites are in all religions the most 
stable aiid_ lasting element, and consequently that in which we 
can best discover the spirit of ancient cults. ^ 

That great pioneer of the science of religion, Robertson Smith, 
took this view. He ^^ rote as follows: 

In connection with every religion, whether ancient or modern, 
we find on the one hand certain behefs, and on the other certain in- 
stitutions, ritual practices and rules of conduct. Our modem habit 
is to look at religion from the side of belief rather than that of practice; 
for, down to comparatively recent times, almost the only forms of religion 
seriously studied in Europe have been those of the various Christian 
Churches, and all parts of Christendom are agreed that ritual is important 

^ 'Les rites 6tant dans toutes les religions I'^l^ment le plus consistant et le 
plus durable, celui, par consequent, ou se d^couvre le mieux I'esprit des cultes 
anciens.' — Essai historique sur le Sacrifice, Paris, 1920, p. i. 


only in connection with its interpretation. Thus the study of religion has 
meant mainly the study of Christian beliefs, and instruction in religion 
has habitually begun with the creed, religious duties being presented 
to the learner as flowing from the dogmatic truths he is taught to accept. 
All this seems to us so much a matter of course that, when we approach 
some strange or antique religion, we naturally assume that here also 
our first business is to search for a creed, and find in it the key to ritual 
and practice. But the_anjiqiie religions^ad for the most part no creed; 
they consisted entirely of institutions and practices. No doubt men will 
not habitually follow certain practices without attaching a meaning to 
them; but as a rule we find that while the practice was rigorously fixed, 
the meaning_attached to it was extremely vague, and the same rite was 
explained by different people Jn different ways, without any question of 
orthodoxy or heterodoxy arising in consequence. In ancient Greece, for 
example, certain things were done at a temple, and people were agreed that 
it would be impious not to do them. But if you asked why they were done 
you would probably have had several mutually contradictory explanations 
from different persons, and no one would have thought it a matter of the 
least religious importance which of these you chose to adopt. Indeed, the 
explanations offered would not have been of a kind to stir any strong feel- 
ing; for in most cases they would have been merely different stories as to 
the circumstances under which the rite first came to be established, by the 
command or by the direct example of the god. The rite, in short, was 
connected not with dogma but with a myth.^ 

... It is of the first importance to realise clearly from the outset 
that ritual and practical usage were, strictly speaking, the sum-total of 
ancient religions. Religion in primitive times was not a system of belief 
with practical applications; it was a body of fixed traditional practices 
to which every member of society conformed as a matter of course. Men 
would not be men if they agreed to do certain things without having a 
reason for their action; but in ancient religion the reason was not first 
formulated as a doctrine and then expressed in practice, but conversely, 
practice preceded doctrinal theory. Men form general rules of conduct 
before they begin to express general principles in words; political in- 
stitutions are older than political theories, and in like manner religious 
institutions are older than religious theories. This analogy is not ar- 
bitrarily chosen, for in fact the parallelism in ancient society between 
religious and political institutions is complete. In each sphere great 
importance was attached to form and precedent, but the explanation 
why the precedent was followed consisted merely of a legend as to its 
first establisliment. That the precedent, once established, was authoritative 
did not appear to require any proof. The rules of society were based on 
precedent, and the continued existence of the society was sufficient 
reason why a precedent once set should continue to be followed.* 

^ W. Robertson Smith, Lectures on the Religion of the Semites, 1907, pp. 16-17. 
* op. cit., p. 20. 


The relative stability of rites and the variability of doctrines 
can be illustrated from the Christian religions. The two essential 
rites of all Christian religions are baptism and the eucharist, and 
we know that the latter solemn sacrament is interpreted differently 
in the Orthodox Church, the Roman Church and the Anglican 
Church. The modern emphasis on the exact formulation of 
beliefs connected with the rites rather than on the rites themselves 
is demonstrated in the way in w^hich Christians have fought with 
and killed one another over differences of doctrine. 

Thirty-seven years ago (1908), in a fellowship thesis on the 
Andaman Islanders (which did not appear in print till 1922), 
I formulated briefly a general theory of the social function of rites 
and ceremonies. It is the same theory that underlies the remarks I 
shall offer on this occasion. Stated in the simplest possible terms 
the theory is that an orderly social life amongst human beings 
depends upon the presence in the minds of the members of a 
society of certain sentiments, which control the behaviour of the 
individual in his relation to others. Rites can be seen to be the 
regulated symbolic expressions of certain sentiments. Rites cart 
therefore be shown to have a specific social function when, and 
to the extent that, they have for their effect to regulate, maintain 
and transmit from one generation to another sentiments on which 
the constitution of the society depends. I ventured to suggest as a 
general formula that religionjs.Xia:ywJiejre_an expression in one 
form or anQther_Df__^a sense of dependence^ on a po wder outside 
ourselYes,^^-^Gwer_:which~\ve^rhay speajc^of as a spiritual or moralj 

This theory is by no means new. It is to be found in the writings 
of the philosophers of ancient China. It is most explicit in the 
teachings of Hsiin TziQ who lived in the third century' B.C., and 
in the Book of Rites (the Li Chi), which was compiled some time 
later. The Chinese wTiters do not write about religion. I am doubt- 
ful if there is in Chinese any w^ord which will convey just what 
we understand by the word religion. They write about //', and the 
word is variously translated as ceremonial, customary' moralit}'', 
rites, rules of good manners, propriety. But the character by which 
this word is written consists of two parts, of which one refers 
to spirits, sacrifice and prayer, and the other originally meant a 
vessel used in performing sacrifices. We may therefore appro- 
priately translate li as 'ritual'. In any case what the ancient 


philosophers are chiefly concerned with are the rites of mourning 
and sacrificial rites. 

There is no doubt that in China, as elsewhere, it \A'as thought 
that many or all of the religious rites were efficacious in the sense 
of averting evils and bringing blessings. It was believed that the 
seasons would not follow one another in due order unless the 
Emperor, the Son of Heaven, performed the established rites at 
the appropriate times. Even under the Republic a reluctant magis- 
trate of a hsien may be compelled by public opinion to take the 
leading part in a ceremony to bring rain. But there developed 
among the scholars an attitude which might perhaps be called 
rationalistic and agnostic. For the most part the question of the 
efficacy of rites was not considered. What was thought important 
was the social function of the rites, i.e. their effects in producing 
and maintaining an orderly human society. 

In a text that is earlier than Confucius we read that 'sacrifice 
is that through which one can show one's filial piety and give 
peace to the people, pacify the country and make the people 
settled. ... It is through the sacrifices that the unity of the 
people is strengthened' {CKu Yii, II, 2). 

You know that one of the major points of the teaching of 
Confucius was the importance of the proper performance of rites. 
But it is said of Confucius that he would not discuss the super- 
natural. ^ In the Confucian philosophy, music and ritual are con- 
sidered as means for the establishment and preservation of social 
order, and regarded as superior to laws and punishments as means 
to this end. We take a very different view of music, but I may 
remind you that Plato held somewhat similar ideas, and I suggest 
that an anthropological study of the relations between music (and 
dancing) and religious rituals would provide some interesting 
results. In the Book of Rites one section (the Yiieh Chi) is con- 
cerned with music. The third paragraph reads: 

The ancient kings were watchful in regard to the things by which the 
mind was affected. And so they instituted ceremonies to direct men's 
aims aright; music to give harmony to their voices; laws to unify their 
conduct; and punishments to guard against their tendencies to evil. 
The end to which ceremonies, music, punishments and laws conduct 

^ Analects, VII, 20. Waley translates this passage as: 'The Master never 
talked of prodigies, feats of strength, disorders or spirits.' 


is one; they are the instruments by which the minds of the people are 
assimilated, and good order in government is made to appear. ^ 

The view of religion that we are here concerned with might 
be summed up in the following sentence from the Book of Rites, 
'Ceremonies are the bond that holds the multitudes together, and 
if the bond be removed, those multitudes fall into confusion.' 

The later Confucian philosophers, beginning with Hsiin 
Tzu, paid attention to the ways in which rites, particularly the 
mourning and sacrificial rites, performed their function of main- 
taining social order. The chief point of their theory is that the rites 
serve to 'regulate' and 'refine' human feelings. Hsiin Tzii says: 

Sacrificial rites are the expressions of man's aflfectionate longings. 
They represent the height of altruism, faithfulness, love and reverence. 
They represent the completion of propriety and refinement. ^ 

Of the mourning rites Hsiin Tzu says: 

The rites (li) consist in being careful about the treatment of life and 
death. Life is the beginning of man, Death is the end of man. When 
the end and begirming are both good, the way of humanity is complete. 
Hence the Superior Man respects the beginning and venerates the end. 
To make the end and beginning uniform is the practice of the Superior 
man, and is that in which lies the beauty of li and standards of justice (0- 
For to pay over-attention to the living and belittle the dead would be to 
respect them when they have knowledge and disrespect them when they 
have not. . . . 

The way of death is this: once dead, a person cannot return again. 
[It is in realising this that] the minister most completely fulfils the honour 
due to his ruler, and the son the honour of his parents. 

Funeral rites are for the living to give beautified ceremonial to the 
dead; to send off the dead as if they were living; to render the same service 
to the dead as to the living; to the absent as to the present; and to make 
the end be the same as the beginning. . . 

Articles used in life are prepared so as to be put into the grave, as if 
[the deceased] were only moving house. Only a few things are taken, 
not all of them. They are to give the appearance, but are not for prac- 
tical use. . . . Hence the things [such as were used] in life are adorned, 
but not completed, and the 'spiritual utensils' are for appearance but 
not use. . . .^ 

^ Legge's translation. An alternative translation of the last sentence would 
be: 'Rites, music, punishments, laws have one and the same end, to unite 
hearts and establish order.' 

^ The translations from Hsiin Tzu are those of Fung Yu Lan and are quoted 
from his History of Chinese Philosophy, Peiping, 1937. 

^ Fung Yu Lan translates by the term 'spiritual utensils' the Chinese ming 
ch'i, which Legge in the following passage from the Book of Rites translates as 


Hence the funeral rites are for no other purpose than to make clear 
the meaning of death and life, to send off the dead with sorrow and 
reverence, and when the end comes, to prepare for storing the body 
away. . . . Service to the living is beautifying their beginning; sending 
off the dead is beautifying their end. When the end and the beginning 
are both attended to, the service of the filial son is ended and the way of 
the Sage is completed. Slighting the dead and over-emphasising the 
living is the way of Mo (Tzfi).^ Shghting the living and over-attention 
to-the- dead is the way of superstitionT^KiTIing the living to send off the 
dead is murder.^ The method and manner of li and standards of justice (i) 
is to send off the dead as if they were alive, so that in death and life, the 
end and the beginning, there is nothing that is not appropriate and good. 
The Confucian does this. 

The view taken by this school of ancient philosophers was 
that ^rehgious rites have important social functions which are 
independent of any behefs that may be held as to the efficacy 
of the rites. The rites gave regulated expression to certain human 
feelings and sentiments and so kept these sentiments alive and 
active. In turn it was these sentiments which, by their control 
of or influence on the conduct of individuals, made possible the 
existence and continuance of an orderly social life.'^ 

It is this theory that I propose for your consideration. Applied, 
not to a single society such as ancient China, but to all human 
societies, it points to the correlation and co-variation of difl"erent 
characteristics or elements of social system4!.Societies differ from 
one another in their structure and constitution and therefore 
in the customary rules of behaviour of persons one to another. 
The system of sentiments on which the social constitution depends 
must therefore vary in correspondence with the difference of 
constitution. In so far as religion has the kind of social function 

'vessels to the eye of fancy': 'Confucius said, "In dealing with the dead, if we 
treat them as if they were entirely dead, that would show a want of affection, 
and should not be done; or, if we treat them as if they were entirely alive, 
that would show a want of wisdom, and should not be done. On this account 
the vessels of bamboo [used in connection with the burial of the dead] are not 
fit for actual use; those of earthenware cannot be used to wash in; those of wood 
are incapable of being carved; the lutes are strung, but not evenly; the pandean 
pipes are complete, but not in tune; the bells and musical stones are there, 
but they have no stands. They are called vessels to the eye of fancy; that is 
[the dead] are thus treated as if they were spiritual intelligencies." ' Legge, 
The Sacred Books of China, Part III, The Li Ki, I-X, Oxford, 1885, p. 148. 

^ Mo Tzu was a philosopher who criticised the mourning rites as being 

* Referring to the ancient practice of human sacrifice at the burial of im- 
portant persons. 


that the theory suggests, relmmn piust also yar^Mn^coiresponjenr.e 
with the manner in which the society is constituted. In a social 
system constituted on the basis of nations which make war on one 
another, or stand ready to do so, a well-developed sentiment of 
patriotism in its members is essential to maintain a strong nation. 
In such circumstances patriotism or national feeling may be 
given support by religion. Thus the Children of Israel, when they 
invaded the land of Canaan under the leadership of Joshua, were 
inspired by the religion that had been taught to them by Moses 
and was centred upon the Holy Tabernacle and its rites. 

War or the envisaged possibility of war is an essential element 
in the constitution of great numbers of human societies, though 
the warlike spirit varies very much from one to another. It is 
thus in accordance with our theory that one of the social functions 
of religion is in connection with war. Jt can give men faith and 
confidence and devotion when they go out to do battle, whether 
they are the aggressors or are resisting aggression. In the recent 
conflict the German people seem to have prayed to God for 
victory not less fervently than the people of the allied nations. 

It will be evident that to test our theor}' we must examine many 
societies to see if there is a demonstrable correspondence of the 
religion or religions of any one of them and the manner in which 
that society is constituted. If such a correspondence can be made 
out, w^e must then try to discover and as far as possible define 
the major sentiments that find their expression in the religion 
and at the same time contribute to the maintenance of stability 
in the society as constituted. 

An important contribution to our study is to be found in a 
book that is undeservedly neglected by anthropologists, La Citi 
antique, by the historian Fustel de Coulanges. It is true that it 
was written some time ago (1864) and that in some matters it 
may need correction in the fight of later historical research, but 
it remains a valuable contribution to the theory of the social 
function of religion. 

The purpose of the book is to show the point-by-point 
correspondence between religion and the constitution of society 
in ancient Greece and Rome, and how in the course of history 
the two changed together. It is true that the author, in conformity 
with the ideas of the nineteenth century, conceived this cor- 
relation between two sets of social features in terms of cause and 


effect, those of one set being thought of as the cause producing those 
of the other set. The men of the ancient world, so the argument 
runs, came to hold certain beliefs about the souls of the dead. 
As the result of their beliefs they made offerings at their tombs. 

Since the dead had need of food and drink it appeared to be a duty 
of the living to satisfy this need. The care of supplying the dead with 
sustenance was not left to the caprice or to the variable sentiments of 
of men; it was obligatory. Thus a complete religion of the dead was 
established, whose dogmas might soon be effaced, but whose rites 
endured until the triumph of Christianity. ^ 

It was a result of this religion that ancient society came to be 
constituted on the basis of the family, the agnatic lineage and the 
gens, with its laws of succession, property, authority and marriage. 

A comparison of beliefs and laws shows that a primitive religion 
constituted the Greek and Roman family, established marriage and 
paternal authority, fixed the order of relationship, and consecrated the 
right of property and the right of inheritance. This same religion, after 
having enlarged and extended the family, formed a still larger association, 
the city, and reigned in that as it had reigned in the family. From it came 
all the institutions, as well as all the private law, of the ancients. It was 
from this that the city received all its principles, its rules, its usages and 
its magistracies. But, in the course of time, this ancient religion became 
modified or effaced, and private law and political institutions were modi- 
fied with it. Then came a series of revolutions, and social changes 
regularly followed the development of knowledge.^ 

In his final paragraph the author writes: 

We have written the history of a belief. It was established and human 
society was constituted. It was modified, and society underwent a series 
of revolutions. It disappeared and society changed its character.^ 

This idea of the primacy of belief and of a causal relation in 
which the religion is the cause and the other institutions are the 
effect is in accordance with a mode of thought that was conmion 
in the nineteenth century. We can, as I indeed do, completely 
reject this theory and yet retain as a valuable and permanent 
contribution to our subject a great deal of what Fustel de Coulanges 
wrote. We can say that he has produced evidence that in ancient 
Greece and Rome the religion on the one side and the many 
important institutions on the other are closely united as inter- 

^ The Ancient City (trans. Willard Small), p. 23. 
^ op cit., p. 12. ^ op cit., p. 529. 


dependent parts of a coherent and unified system. The rehgion 
was an essential part of the constitution of the society. The form of 
the rehgion and the form of the social structure correspond one 
with the other. We cannot, as Fustel de Coulanges says, under- 
stand the social, juridical and political institutions of the ancient 
societies unless we take the religion into account. But it is equally 
true that we cannot understand the religion except by an examina- 
tion of its relation to the institutions. 

A most important part of the religion of ancient Greece and 
Rome was the worship of ancestors. We may regard this as one 
instance of a certain type of religion. A religious cult of the same 
general kind has existed in China from ancient times to the present 
day. Cults of the same kind exist to-day and can be studied in 
many parts of Africa and Asia. It is therefore possible to make a 
wide comparative study of this type of religion. In my own ex- 
perience it is in ancestor- worship that we can most easily discover 
and demonstrate the social function of a religious cult. 

The term 'ancestor- worship' is sometimes used in a wide, 
loose sense to refer to any sort of rites referring to dead persons. 
I propose to use it in a more limited and more precisely defined 
sense. The cult group in this religion consists solely of persons 
related to one another by descent in one line from the same 
ancestor or ancestors. In most instances descent is patrilineal, 
through males. But in some societies, such as the Bakongo in 
Africa and the Nayar in India, descent is matrilineal, and the 
cult group consists of descendants of a single ancestress. The rites 
in which the members of the group, and only they, participate 
have reference to their own ancestors, and normally they include 
the making of offerings or sacrifices to them. 

A particular lineage consists of three or more generations. 
A lineage of four or five generations will normally be included as a 
part in one of six or seven generations. In a well-developed system 
related lineages are united into a larger body, such as the Roman 
gens, or what may be called the clan in China. In parts of China 
we can find a large body of persons, numbering in some instances 
as much as a thousand, all having the same name and tracing their 
descent in the male line from a single ancestor, the founder of 
the clan. The clan itself is divided into lineages. 

A lineage, if it is of more than tliree or four generations, 
includes both living persons and dead persons. What is called 


ancestor-worship consists of rites carried out by members of a 
larger or smaller lineage (i.e. one consisting of more or fewer 
generations) with reference to the deceased members of the 
lineage. Such rites include the making of offerings, usually of 
food and drink, and such offerings are sometimes interpreted as 
the sharing of a meal by the dead and the living. 

In such a society, what gives stability to the social structure 
is the solidarity and continuity of the hneage, and of the wider 
group (the clan) composed of related lineages. For the individual, 
his primary duties are those to his lineage. These include duties 
to the members now living, but also to those who have died and to 
those who are not yet born. In the carrying out of these duties 
he is controlled and inspired by the complex system of sentiments 
of which we may say that the object on which they are centred 
is the lineage itself, past, present and future. It is primarily this 
system of sentiments that is expressed in the rites of the cult of the 
ancestors. The social fujiction of the rites is obvious: by giving 
solemn and collective expression to them the rites reaffirm, renew 
and strengthen those sentiments on which the social solidarity 

We have no means of studying how an ancestor-worshipping 
society comes into existence, but we can study the decay of this 
type of system in the past and in the present. Fustel de Coulanges 
deals with this in ancient Greece and Rome. It can be observed 
at the present time in various parts of the world. The scanty 
information I have been able to gather suggests that the lineage 
and joint-family organisation of some parts of India is losing 
something of its former strength and solidarity and that what we 
should expect as the inevitable accompaniment of this, a weaken- 
ing of the cult of ancestors, is also taking place. I can speak with 
more assurance about some African societies, particularly those 
of South Africa. The effect of the impact of European culture, 
including the teaching of the Christian missionaries, is to weaken 
in some individuals the sentiments that attach them to their 
lineage. The disintegration of the social structure and the decay 
of the ancestral cult proceed together. 

Thus for one particular type of religion I am ready to affirm 
that the general theory of the social function of religions can be 
fully demonstrated. 

A most important contribution to our subject is a work of 


Emile Durkheim published in 1912. The title is Les Formes 
elementaires de la Vie religieuse, but the sub-title reads: La Systeme 
totemique en Australie. It is worth while mentioning that Durkheim 
was a pupil of FusteLde Coulanges at the Ecole Normale Superieure 
and that he himself said that the most important influence on the 
development of his ideas about religion was that of Robertson 

Durkheim's aim was to establish a general theory of the nature 
of religion. Instead of a wide comparative study of many religions, 
he preferred to take a simple type of society and carry out an in- 
tensive and detailed analysis, and for this purpose he selected the 
aboriginal tribes^of_Australia. He held the view that these tribes 
represent the simplest type of society surviving to our own times, 
but the value of his analysis^tSTTTTTorWay affected if we refuse to 
accept this view, as I do myself. 

The value of Durkheim's book is as an exposition of a general 
theory of religion which had been developed with the collabora- 
tion of Henri Hubert and Marcel Mauss, starting from the founda- 
tions provided by Robertson Smith. Durkheim's exposition of this 
theory has often been very much misunderstood. A clear, though 
very brief, statement of it is to be found in the Introduction 
written by Henri Hubert in 1904 for the French translation of the 
Manuel d'Histoire des Religions of Chantepie de la Saussaye. 
But it is not possible on this occasion to discuss this general 
theory. I wish only to deal with one part of Durkheim's work, 
namely his 'theory that religious ritual is an expxsssion of the 
unity of soaej^Tand that its function is to 're-create' the society 
of the social order by reaffirming and strengthening the senti- 
ments on which the social solidarity and therefore the social order 
itself depend.^ This theory he tests by an examination of the 
totemic ritual of the Australians. For while Frazer regarded the 
totemic rites of the Australian tribes as being a matter of magic, 
Durkheim treats them as religious because the rites themselves 
are sacred and have reference to sacred beings, sacred places and 
sacred objects. 

In 191 2 very much less was known about the Australian 
aborigines than is known at present. Some of the sources used 
by Durkheim have proved to be unreliable. The one tribe that 
was well known, through the writings of Spencer and Gillen and 

^ op cit., pp. 323, 497 and elsewhere. 


Strehlow — the Aranda — is in some respects atypical. The in- 
formation that Durkheim could use was therefore decidedly 
imperfect. Moreover, it cannot be said that his handling of this 
material was all that it might have been. Consequently there are 
many points in his exposition wliich I find unacceptable. Never- 
theless, I think that Durkheim's major thesis as to the social 
function of the totemic rites is valid and only requires revision 
and correction in the light of the more extensive and more exact 
knowledge we now have.^ 

The beings to which the Australian cult refers are commonly 
spoken of as 'totemic ancestors', and I have myself used the term. 
But it is somewhat misleading, since thej_arejnythical beings and ' 
not ac.cest.9rs in the same sense as the dead persons conmiemorated 
in ancestor-worship. In the cosmology of the Australian natives 
the cosmos, the ordered universe, including both the order of 
nature and the social order, came into existence at a time in the 
past which I propose to speak of as the World- Dawn, for this 
name corresponds to certain ideas that I have found amongst the 
aborigines of some tribes. This order (of nature and of society) 
resulted from the doings and adventures of certain sacred beings. 
These beings, whom I shall call the Dawn Beings, are the totemic 
ancestors of ethnological literature. The explanations of topo- 
graphical features, of natural species and their characteristics, and 
of social laws, customs and usages are given in the form of myths 
about the happenings of the World- Dawn. 

The cosmos is ruled by law. But whereas we think of the 
laws of nature as statements of what invariably does happen 
(except, of course, in miracles), and of moral or social laws as what 
ought to be observed but are sometimes broken, the Australian 
does not make this distinction. For him men and women ought to 
observe the rules of behaviour that were fixed for all time by the 
events of the World- Dawn, and similarly the rain ought to fall in 
its proper season, plants should grow and produce fruit or seed, 
and animals should bear young. But there are irregularities in 
human society and in nature. 

In what I shall venture to call the totemic religion of the 
Australian aborigines, there are two main types of ritual. One of 
these consists of rites carried out at certain spots which are com- 

^ For a criticism of some points in Durkheim's work, see 'The Sociological 
Theory of Totemism' in this volume. 


monly referred to as 'totem centres'. A totem centre is a spot that is 
specially connected with some species of object, most commonly 
with a particular species of animal or plant, or with an aspect of 
nature such as rain or hot weather. Each centre is associated with 
one (or occasionally more than one) of the Dawn Beings. Fre- 
quently the Being is said to have gone into the ground at this spot. 
For each totem centre there is a myth connecting it w ith the events 
of the World-Dawn. The totem centre, the myth connected with it 
and the rites that are performed~there, belong to the local group 
that owns the territory within which the totem centre lies. Each 
totem centre is thought of as containing, in a rock or a tree or a 
pool of water or a heap of stones, what we may perhaps call the 
hfe-spirit or life-force of the totem species. 

The rites performed at the totem centre by the members of the 
local group to which it belongs, or under their leadership and 
direction, are thought tojenew the vitality of this life-spirit of the 
species. In eastern Australia theTofem centre is spoken of as the 
'home' or 'dwelling-place' of the species, and the rites are called 
'stirring up'. Thus, the rite at a rain totem centre brings the rain 
in its due season, that at a kangaroo totem centre ensures the 
supply of kangaroos, and that at the baby totem centre provides 
for the birth of children in the tribe. 

These rites imply a certain conception, which I think we can 
call specifically a religious conception, of the place of man in the 
universe. Man is dependent upon what we call nature: on the 
regular successions of the seasons, on the rain falling when it should, 
on the growth of plants and the continuance of animal life. But, 
as I have already said, while for us the order of nature is one . 
thing and the social order another, for the Australian they are 1 
two parts of a single order. Well-being, for the individual or ' 
for the society, depends on the continuance of this order free 
from serious disturbance. The Australians believe that they can \ 
ensure this continuance, or at least contribute to it, by their actions, 
including the regular performance of the totemic rites. 

In the rites that have been described, each group takes care 
(if we may so express it) of only a small part of nature, of those 
few species for which it owns totem centres. The preservation of 
the natural order as a whole therefore depends on the actions of 
many dirferent groups. 

The social structure of the Australian natives is based on two 


things: a system of local groups, and a system of kinship based on 
the family. Each small local group is a closed patrilineal descent 
group; that is, a man is born into the group of his father and his 
sons belong to his group. Each group is independent and autono- 
mous. The stability and continuity of the social structure depends 
on the strong solidarity of the local group. 

Where there existed the totemic cult which I have just de- 
scribed (and it existed over a very large part of Australia), each 
local group was a cult group. The totemic ritual served to express 
thejmity^ and solidarity of the group" and its individuality and 
separation from other groups by the special relation of the group 
to its sacra: the totem centre or centres, the Dawn Beings associated 
with them, the myths and songs referring to those Beings, and the 
totems or species connected with the centres. This aspect of the 
social function of totemism was emphasised, and I think somewhat 
over-emphasised, by Durkheim. 

There is, however, another aspect, for while the local totemic 
groups are separate individual and continuing social entities, they 
are also part of a wider social structure. This wider structure is 
provided by the kinship system. For an individual in Australian 
native society, every person with whom he has any social contact 
is related to him by some bond of kinship, near or distant, and 
the regulation of social life consists essentially of rules concerning 
behaviour towards different kinds of kin. For example, a man 
stands in very close relation to his mother's local group and, in 
many tribes, in a very close relation to its sacra: its totems, totem 
centres and totemic rites. 

While Australian totemism separates the local groups and 
gives each an individuality of its own, it also links the groups to- 
gether. For while each group is specially connected with certain 
parts of the natural order (e.g. with rain, or with kangaroo) and 
with certain of the Beings of the World- Dawn, the society as 
a whole is related through the totemic religion to the whole 
order of nature and to the World- Dawn as a whole. This is best 
seen in another kind of totemic cult, part of which consists of 
sacred dramas in which the performers impersonate various Dawn 
Beings. Such dramatic dances are only performed at those religious 
meetings at which a number of local groups come together, and 
it is on these occasions that young men are initiated into manhood 
and into the religious life of the society. 


Australian society is not merely a collection of separate 
local groups; it is also a body of persons linked together in the 
kinship system. Australian totemism is a cosmological system by 
which the phenomena of nature are incorporated in the kinship 
organisation. When I v/as beginning my work in Australia in 191 o, 
a native said to me, 'Bungurdi (kangaroo) [is] my kadja (elder 
brother).' This simple sentence of three words gives the clue 
to an understanding of Australian totemism. The speaker did not 
mean that individuals of the kangaroo species are his brothers. 
He meant that to the kangaroo species, conceived as an entity, he 
stood in a social relation analogous to that in which a man stands to 
his elder brother in the kinship system. I am sorry that there is not 
time on this occasion to expound this thesis more fully. 

The account I have just given of Australian totemism differs 
considerably from that given by Durkheim. But far from con- 
tradicting, it confirms Durkheim's fundamental general theory 
as to the social function of the totemic religion of Australia and 
its rites. The two kinds of totemic cult are the demonstration, in 
symbolic action, of the structure of Australian society and its 
foundations in a mythical and sacred past. In maintaining the 
social cohesion and equilibrium, the religion plays a most im- 
portant part. The rehgion is an intrinsic part of the constitution 
of society. 

I have dwelt, if only cursorily, with two types of religion: 
ancestor- worship and Australian totemism. In both of them it is 
possible to demonstrate the close correspondence of the form of 
religion and the form of the social structure. In both it is possible 
to see how the religious rites reaffirm and strengthen the senti- 
ments on which the social order depends. Here then are results 
of some significance for our problem. They point to a certain line 
of investigation. We can and should examine other religions in the 
light of the results already reached. But to do this we must study 
religions in action; we must try to discover the effects of active 
participation in a particular cult, first the direct effects on the 
individual and then the further effects on the society of which these 
individuals are members. When we have a sufficient number of 
such studies, it will be possible to establish a general theory of the 
nature of religions and their role in social development. 

In elaborating such a general theory it will be necessary to 
determine by means of comparative studies the relations between 


religion and morality. There is only time to refer very briefly 
here to the question of religion and morality. As representing a 
theory that seems to be widely held, I quote the following passages 
from Tylor: 

One great element of religion, that moral element which among the 
higher nations forms its most vital part, is indeed little represented in the 
religion of the lower races. ^ 

The comparison of savage and civilised religions brings into view, 
by the side of a deep-lying resemblance in their philosophy, a deep-lying 
contrast in their practical action on human life. So far as savage religion 
can stand as representing natural religion, the popular idea that the moral 
government of the universe is an essential tenet of natural religion simply 
falls to the ground. Savage animism is almost devoid of that ethical 
element which to the educated modern mind is the very mainspring of 
practical religion. Not, as I have said, that morality is absent from the 
life of the lower races. Without a code of morals, the very existence of 
the rudest tribe would be impossible; and indeed the moral standards 
of even savage races are to no small extent well-defined and praiseworthy. 
But these ethical laws stand on their own ground of tradition and public 
opinion, comparatively independent of the animistic beliefs and rites 
which exist beside them. The lower animism is not immoral, it is 
unmoral. . . . The general problem of the relation of morality to religion 
is difficult, intricate, and requiring immense array of evidence.^ 

I agree with Tylor that the problem of the relation of morality 
to religion is difficult and intricate. But I wish to question the 
validity of the distinction he makes between the religions of 
savage sand those of civilised peoples, and of his statement that 
the moral element 'is little represented in^the religion of the lower 
races'. I suspect that when this view is held it often means only 
that in the 'lower races' the religion is not associated with the 
kind of morality which exists in contemporary Western societies. 
But societies diifer in their systems of morals as in other aspects 
of the social system, and what we have to examine in any given 
society is the relation of the religion or religions of that society 
to their particular system of moraUty. 

Dr. R. F. Fortune, in his book on Manus religion, has chal- 
lenged the dictum of Tylor. ^ The religion of Manus is what may 
be called a kind of spiritualism, but it is not ancestor- worship in 

^ Tylor, Primitive Culture, 3rd ed., 1891, Vol. I, p. 427. 

- op cit., Vol. II, p. 360. 

' R. F. Fortune, Manus Religion, Philadelphia, 1935, pp. s and 356. Dr. 
Fortune's book is a useful contribution to the study of the social function of 
religion and deals with a religion of a very unusual type. 


the sense in which I have used the term in this lecture. The 
Manus code of morals rigidly forbids sexual intercourse except 
between husband and wife, condemns dishonesty and insists on 
the conscientious fulfilment of obligations, including economic 
obligations, towards one's relatives and others. Offences against 
the moral code bring down on the offender, or on his household, 
punishment from the spirits, and the remedy is to be found in 
confession and reparation for wrong. 

Let us now reconsider the case of ancestor-worship. In the 
societies which practise it, the most important part of the moral 
code is that which concerns the conduct of the individual in 
relation to his lineage and clan and the individual members thereof. 
In the more usual form of ancestor- worship, infractions of this 
code fall under religious or supernatural sanctions, for they 
are offences against the ancestors, who are believed to send 

Again we may take as an example of the lower races the 
aborigines of Australia. Since the fundamental social structure is 
a complex system of widely extended recognition of relations of 
kinship, the most important part of the moral code consists of 
the rules of behaviour towards kin of different categories. One of 
the most immoral actions of which a man can be guilty is having 
sexual relations with any woman who does not belong to that 
category of his kinsfolk into w^hich he may legally marry. 

The moral law of the tribe is taught to young men in the very 
sacred ceremonies known as initiation ceremonies. I will deal only 
with the Bora ceremonies, as they are called, of some of the tribes 
of New South Wales. These ceremonies were instituted in the 
time of the World-Da\Mi by Baiame, who killed his own son 
Daramulun (sometimes identified with the sacred bull-roarer) and 
on the third day brought him back to life. As the ceremony is 
conducted, the initiates all 'die' and are brought back to life on 
the third day.^ 

On the sacred ceremonial ground where these initiations take 
place there is usually an image of Baiame made of earth, and 
sometimes one of Baiame's wife. Beside these images sacred rites 

^ The suggestion has been made that we have here the influence of Christian- 
ity, but that opinion can be dismissed. The idea of ritual death and rebirth is 
very widespread in religion, and the three-day period is exemplified every 
month in every part of the world by the death and resurrection of the moon. 


are shown to the initiates, and sacred myths about Baiame are 

Now Baiame instituted not only the initiation ceremonies, 
which are, amongst other things, schools of morals for young men, 
but also the kinship system with its rules about marriage and 
behaviour towards different categories of kin. To the question, 
'Why do you observe these complex rules about marriage?' 
the usual answer is, 'Because Baiame established them'. Thus 
Baiame is the divine law-giver, or, by an alternative mode of 
expression, he is the personification of the tribal laws of 

I agree with Andrew Lang and Father Schmidt that Baiame 
thus closely resembles one aspect of the God of the Hebrews. 
But Baiame gives no assistance in war as Jehovah did for the 
children of Israel, nor is Baiame the ruler or controller of nature, 
of storms and seasons. That position is held by another deity, the 
Rainbow- Serpent, whose image in earth also appears on the 
sacred ceremonial ground. The position held by Baiame is that 
of the Divine Being who established the most important rules of 
morality and the sacred ceremonies of initiation. 

These few examples will perhaps suffice to show that the idea 
that it is only the higher religions that are specially concerned 
with morahty, and that the moral element is little represented 
in the religions of the lower races, is decidedly open to question. 
If there were time I could provide instances from other parts of the 

What makes these problems complex is the fact that law, 
morality and religion are three ways of controlling human conduct 
which in different types of society supplement one another, and 
are combined, in different ways. For the law there are legal 
sanctions, for morality there are the sanctions of public opinion 
and of conscience, for religion there are religious sanctions. 
A single wrongful deed may fall under two or three sanctions. 
Blasphemy and sacrilege are sins and so subject to religious 
sanctions; but they may also sometimes be punished by law as 
crimes. In our own society murder is immoral; it is also a 
crime punishable by death ; and it is also a sin against God, so 
that the murderer, after his sudden exit from this life at the 
hands of the executioner, must face an eternity of torment in 
the fires of Hell. 


Legal sanctions may be brought into action in instances 
where there is no question of morahty or immoraHty, and the 
same is true of rehgious sanctions. It is held by some of the Fathers 
or doctors of the Christian churches that an upright and virtuous 
life devoted to good works will not save a man from Hell unless 
he has attained grace by accepting as true the specific doctrines 
taught by a church, 

Xi^f'-reare different kinds of religious sanctions. The penalty 
for sin may be conceived simply as alienation from God. Or 
there may be a belief in rewards and punishments in an after-life. 
But the most widespread form of the religious sanction is the 
belief that certain actions produce in an individual or in a com- 
munity a condition of ritual pollution, or uncleanness, from which 
it is necessary to be purified. Pollution may result from things 
done unintentionally and unwittingly, as you may see from the 
fifth chapter of the Book of Leviticus. One who unwittingly has 
touched any unclean thing, such as the carcase of an unclean beast, 
is guilty and has sinned and must bear his iniquity. He must make 
a sacrifice, a trespass offering, by which he may be cleansed from 
his sin. 

Ritual uncleanness does not in itself involve moral con- 
demnation. We read in the twelfth chapter of the same Book of 
Leviticus that the Lord instructed Moses that a woman who has 
borne a male child shall be unclean for seven days and her 
purification must continue for a further three and thirty days, 
during which she shall touch no hallowed thing, nor come into the 
sanctuary. If the child she bears is female, the first period of 
uncleanness is to be two weeks and the period of purification 
threescore-and-six days. Thus, it is polluting, but no one can 
suppose that it is immoral, to bear a child, and more polluting 
if the child is female than if it is male. 

The opposite of pollution or sinfulness is holiness. But 
holiness comes not from leading an honest and upright life, but 
from religious exercises, prayer and fasting, the performance of 
penance, meditation and the reading of sacred books. In Hinduism 
the son of a Brahmin is bom holy; the son of a leather-worker is 
born unclean. 

The field covered by morality and that covered by religion 
are dift'erent; but either in primitive or in civilised societies there 
may be a region in which they overlap. 



To return to our main topic, a writer who has dealt with 
the social function of religions on the basis of a comparative 
study is Loisy, who devotes to the subject a few pages of the con- 
cluding chapter of his valuable Essai historique sur le Sacrificed 
Although he differs from Durkheim in some matters, his funda- 
mental theory is, if not identical, at any rate very similar to that 
of the earlier writer. Speaking of what he calls the sacred action 
{V action sacree), of which the most characteristic form is the rite of 
sacrifice, he writes: 

We have seen its rl6e in human societies, of which it has maintained 
and strengthened the social bonds, if indeed it has not contributed in a 
large measure to creating them. It was, in certain respects, the expression 
of them; but man is so made that he becomes more firmly fixed in his 
sentiments by expressing them. The sacred action was the expression 
of social life, of social aspirations, it has necessarily been a factor of 
society. . . . 

Before we condemn out of hand the mirage of religion and the ap- 
paratus of sacrifice as a simple waste of social resources and forces, it is 
proper to observe that, religion having been the form of social conscience, 
and sacrifice the expression of this conscience, the loss was compensated 
by a gain, and that, so far as purely material losses are concerned, there 
is really no occasion to dwell on them. Moreover the kind of sacred 
contribution that was required, without real utility as to the effect that 
was expected from it, was an intrinsic part of the system of renuncia- 
tions, of contributions which, in every human society, are the condition 
of its equilibrium and its conservation. ^ 

But besides this definition of the social function in terms 
of social cohesion and continuity, Loisy seeks for what he calls a 
general formula {formule generale) in which to sum up the part 
that religion has played in human life. Such a formula is useful so 
long as we remember that it is only a formula. The one that 
Loisy offers is that magic and religion have served to give men 

In the most primitive societies it is magic that gives man 
confidence in face of the difficulties and uncertainties, the real and 
imaginary dangers with which he is surrounded. 

A la merci des elements, des saisons, de ce que la terre lui donne ou 
lui refuse, des bonnes ou des mauvaises chances de sa chasse ou de sa 
peche, aussi du hasard de ses combats avec ses semblables, il croit trouver 
le moyen de regulariser par des simulacres d'action ces chances plus ou 

^ 1920, pp. 531-40. ^ op cit., pp. 535-7. 


moins incertaines. Ce qu'il fait ne sert a rien par rapport au but qu'il se 
propose, mais il prend confiance en ses entreprises et en lui-meme, 
il ose, et c'est en osant que reellement il obtient plus ou moins ce qu'il 
veut. Confiance rudimentaire, et pour une humble vie; mais c'est le 
commencement du courage moral.^ 

This is the same theory that was later developed by IVIalinowski 
in reference to the magical practices of the Trobriand Islanders. 

At a somewhat higher stage of development, 'when the social 
organism has been perfected, when the tribe has become a people, 
and this people has its gods, its religion, it is by this religion itself 
that the strength of the national conscience is measured, and it 
is in the service of national gods that men find a pledge of security 
in the present, of prosperity in the future. The gods are as it were 
the expression of the confidence that the people has in itself; 
but it is in the cult of the gods that this confidence is nourished.'^ 

At a still higher stage of social development, the religions 
which give men a promise of immortality give him thereby an 
assurance which permits him to bear courageously the burdens 
of his present life and face the most onerous obligations. 'It is a 
higher and more moral form of confidence in life. '^ ; 

To me this formula seems unsatisfactory in tliat it lays stress 
on what is only one side of the religious (or magical) attitude. I 
offer as an alternative the formula that religion develops in man- 
kind what may be called a §gnse of dependence. What I mean by 
this can best be explained by an example. In an ancestor-wor- 
shipping tribe of South Africa, a man feels that he is dependent 
on his ancestors. From them he has received his life and the cattle 
that are his inheritance. To them he looks to send him children 
and to multiply his cattle and in other ways to care for his well- 
being. This is one side of the matter; on his ancestors he can 
depend. The other side is the belief that the ancestors watch over 
his conduct, and that if he fails in his duties they will not only 
cease to send him blessings, but will visit him with sickness or . 
some other misfortune. He cannot stand alone and depend only J 
on his own efforts; on his ancestors he must depend. 

We may say that the beUefs of the African ancestor- worshipper 
are illusory and his offerings to his gods really useless; that the 
dead of his lineage do not really send him either blessings or 

' op cit., p. 533. * loc cit. * op cit., p. 534. 


punishments. But the Confucians have shown us that a rehgion 
hke ancestor-worship can be rationaUsed and freed from those 
illusory beliefs that we call superstition. For in the rites of com- 
memoration of the ancestors it is sufficient that the participants 
should express their reverential gratitude to those from whom they 
have received their life, and their sense of duty towards those not 
yet born, to whom they in due course will stand in the position 
of revered ancestors. There still remains the sense of dependence. 
The living depend on those of the past; they have duties to those 
living in the present and to those of the future who will depend 
on them. 

I suggest to you that what makes and keeps a man a social 
animal is not some herd instinct, but the sense of dependence in 
the innumerable forms that it takes. The process of socialisation 
begins on the first day of an infant's life and it has to learn that it 
both can and must depend on its parents. From them it has comfort 
and succour; but it must submit also to their control. What I 
am calling the sense of dependence always has these two sides. 
We can face life and its chances and difficulties with confidence 
when we know that there are powers, forces and events on 
which we can rely, but we must submit to the control of our 
conduct by rules which are imposed. The entirely asocial in- 
dividual would be one who thought that he could be completely 
independent, relying only on himself, asking for no help and 
recognising no duties. 

I have tried to present to you .a theory of the social function of 
religion. This theory has been developed by the work of such men 
as Robertson Smith, Fustel de Coulanges, Durkheim, Loisy. 
It is the theory that has guided my own studies for nearly forty 
years. I have thought it worth while to indicate that it existed 
in embryo in the writings of Chinese philosophers more than 
twenty centuries ago. 

Like any other scientific theory it is provisional, subject to 
revision and modification in the light of future research. It is 
offered as providing what seems likely to be a profitable method 
of investigation. What is needed to test and further elaborate the 
theory is a number of systematic studies of various types of religion 
in relation to social systems in which they occur. 

I will summarise the suggestions I have made: 


1. To understand a particular religion we must study its 
effects. The religion must therefore _be_studied in action. 

2. Since human conduct is in large part controlled or directed 
by what have been called sentiments, conceived as mental 
dispositions, it is necessary to discover as far as possible 
what are the senlkoents that are developed in the individual 
as the result of his participation in a particular rehgious 

3. In the study of any religion we must first^pf all examine 
the specifically religious actions, the ceremonies and the 
collective or individual rites. 

4. The emphasis on belief in specific doctrines which charac- 
terises some modem religions seems to be the result of 
certain social developments in societies of complex structure. 

5. In some societies there is a direct and immediate relation 
between the religion and the social structure. This has been 
illustrated by ancestor-worship and Australian totemism. 
It is also true of what w^e may call national religions, such 
as that of the Hebrews or those of the city states of Greece 
and Rome.^ But where there comes into existence a separate 
independent religious structure by the formation of different 
churches or sects or cult-groups within a people, the 
relation of religion to the total social structure is in many 
respects indirect and not always easy to trace. 

6. As a general formula (for whatever such a formula may be 
worth) it is suggested that what is expressed in all rehgions 
is what I have called the sense of dependence in its double 
aspect, and that it is by constantly maintaining this sense 
of dependence that religions perform their social function. 

^ * . . . among the ancients what formed the bond of every society was a 
worship. Just as a domestic altar held the members of a family grouped about 
it, so the city was the collective group of those who had the same protecting 
deities, and who performed the religious ceremony at the same altar.' Fustel 
de Coulanges, op cit., p. 193. 



THE concept of function applied to human societies is 
based on an analogy between social life and organic life. 
The recognition of the analogy and of some of its implica- 
tions is not new. In the nineteenth century the analogy, the concept 
of function, and the word itself appear frequently in social philo- 
sophy and sociology. So far as I know the first systematic 
formulation of the concept as applying to the strictly scientific 
study of society was that of Emile Durkheim in 1895. [Rigles de la 
Methode Sociologique.) 

Durkheim's definition is that the 'function' of a social in- 
stitution is the correspondence between it and the needs (besoins 
in French) of the social organism. This definition requires some 
elaboration. In the first place, to avoid possible ambiguity and 
in particular the possibility of a teleological interpretation, I 
would like to substitute for the term 'needs' the term 'necessary 
conditions of existence', or, if the term 'need' is used, it is to be 
understood only in this sense. It may be here noted, as a point 
to be returned to, that any attempt to apply this concept of function 
in social science involves the assumption that there are necessary 
conditions of existence for human societies just as there are for 
animal organisms, and that they can be discovered by the proper 
kind of scientific enquiry. 

For the further elucidation of the concept it is convenient to 
use the analogy between social life and organic life. Like all 
analogies it has to be used with care. An animal organism is an 
agglomeration of cells and interstitial fluids arranged in relation 
to one another not as an aggregate but as an integrated living 
whole. For the biochemist, it is a complexly integrated system 
of complex molecules. The system of relations by which these 

^ This paper, which is based on comments that I made on a paper read by 
Dr. Lesser to the American Anthropological Association, is reprinted from 
the American Anthropologist, Vol. XXXVII, p. 3, 1935, where it accompanied 
Dr. Lesser's paper. 



units are related is the organic structure. As the terms are here 
used the organism is not itself the structure; it is a collection of 
units (cells or molecules) arranged in a structure, i.e. in a set of 
relations; the organism has a structure. Two mature animals of the 
same species and sex consist of similar units combined in a similar 
structure. The structure is thus to be defined as a set of relations 
between entities. (The structure of a cell is in the same way aj 
set of relations between complex molecules, and the structure of 
an atom is a set of relations between electrons and protons.) As 
long as it lives the organism preserves a certain continuity of 
structure although it does not preserve the complete identity of 
its constituent parts. It loses some of its constituent molecules 
by respiration or excretion; it takes in others by respiration and 
alimentary absorption. Over a period its constituent cells do not 
remain the same. But the structural arrangement of the constituent 
units does remain similar. The process by which this structural 
continuity of the organism is maintained is called life. The 
life-process consists of the activities and interactions of the 
constituent units of the organism, the cells, and the organs into 
which the cells are united. 

As the word function is here being used the life of an organism 
is conceived as the functioning of its structure. It is through and 
by the continuity of the functioning that the continuity of the 
structure is preserved. If we consider any recurrent part of the 
life-process, such as respiration, digestion, etc., its function 
is the part it plays in, the contribution it makes to, the life of 
the organism as a whole. As the terms are here being used a cell 
or an organ has an activity and that activity has a function. It is 
true that we commonly speak of the secretion of gastric fluid as a 
'function' of the stomach. As the words are here used we should 
say that this is an 'activity' of the stomach, the 'function' of which 
is to change the proteins of food into a form in which these are 
absorbed and distributed by the blood to the tissues.^ We may note 
that the function of a recurrent physiological process is thus a 
correspondence between it and the needs (i.e. necessary conditions 
of existence) of the organism. 

^ The insistence on this precise form of terminology is only for the sake 
of the analogy that is to be drawn. I have no objection to the use of the term 
function in physiology to denote both the activity of an organ and the results 
of that activity in maintaining life. 


If we set out upon a systematic investigation of the nature of 
organisms and organic life there are three sets of problems 
presented to us. (There are, in addition, certain other sets of 
problems concerning aspects or characteristics of organic life with 
\ which we are not here concerned.) One is that of morphology — 
what kinds of organic structures are there, what similarities and 
variations do they show, and how can they be classified? Second 
are the problems of physiology — how, in general, do organic 
structures function, what, therefore, is the nature of the life- 
process? Third are the problems of evolution or development — 
how do new types of organisms come into existence? 

To turn from organic life to social life, if we examine such a 
community as an African or Australian tribe we can recognise 
the existence of a social structure. Individual human beings, the 
essential units in this instance, are connected by a definite set of 
social relations into an integrated whole. The continuity of the 
social structure, like that of an organic structure, is not destroyed 
by changes in the units. Individuals may leave the society, by 
death or otherwise; others may enter it. The continuity of structure 
is maintained by the process of social life, which consists of the 
activities and interactions of the individual human beings and of 
the organised groups into which they are united. The social life 
of the community is here defined as the functioning of the social 
structure. The function of any recurrent activity, such as the 
punishment of a crime, or a funeral ceremony, is the part it plays 
in the social life as a whole and therefore the contribution it makes 
to the maintenance of the structural continuity. 

The concept of function as here defined thus involves the 
notion of a structure consisting of a set of relations amongst unit 
entities, the continuity of the structure being maintained by a life- 
process made up of the activities of the constituent units. 

If, with these concepts in mind, we set out on a systematic 
investigation of the nature of human society and of social life, 
we find presented to us three sets of problems. First, the problems 
of social morphology — ^what kinds of social structures are there, 
what are their similarities and differences, how are they to be 
classified? Second, the problems of social physiolog}'^ — how do 
social structures function? Third, the problems of development — 
how do new types of social structure come into existence? 

Two important points where the analogy between organism and 


society breaks down must be noted. In an animal organism it is 
possible to observe the organic structure to some extent in- 
dependently of its functioning. It is therefore possible to make a 
morphology which is independent of physiology. But in human 
society the social structure as a w hole can only be observed in its 
functioning. Some of the features of social structure, such as 
the geographical distribution of individuals and groups can be 
directly observed, but most of the social relations which in their 
totality constitute the structure, such as relations of father and 
son, buyer and seller, ruler and subject, cannot be observed 
except in the social activities in which the relations are functioning. 
It follows that a social morphology cannot be established in- 
dependently of a social physiology. 

The second point is that an animal organism does not, in the 
course of its life, change its structural type. A pig does not become 
a hippopotamus. (The development of the animal from ger- 
mination to maturity is not a change of type since the process in 
all its stages is typical for the species.) On the other hand a 
society in the course of its history can and does change its structural 
type without any breach of continuity. 

By the definition here offered 'function' is the contribution 
which a partial activity makes to the total activity of which it is a 
part. The function of a particular social usage is the contribution 
it makes to the total social life as the functioning of the total social 
system. Such a view implies that a social system (the total social 
structure of a society together with the totality of social usages 
in which that structure appears and on which it depends for its 
continued existence) has a certain kind of unity, which we may 
speak of as a functional unity. We may define it as a condition 
in which all parts of the social system work together with a 
sufficient degree of harmony or internal consistency, i.e. without 
producing persistent conflicts which can neither be resolved nor 

This idea of the functional unity of a social system is, of course, [ 
a hypothesis. But it is one which, to the functionalist, it seems i 
worth while to test by systematic examination of the facts. 

There is another aspect of functional theory that should be 
briefly mentioned. To return to the analogy of social life and 

^ Opposition, i.e. organised and regulated antagonism, is, of course, an 
essential feature of every social system. 


organic life, we recognise that an organism may function more or 
less efficiently and so we set up a special science of pathology 
to deal with all phenomena of disfunction. We distinguish in an 
organism what we call health and disease. The Greeks of the fifth 
century B.C. thought that one might apply the same notion to 
society, to the city-state, distinguishing conditions of eunomia, 
good order, social health, from dysnomia, disorder, social ill- 
health. In the nineteenth century Durkheim, in his application 
of the notion of function, sought to lay the basis for a scientific 
social pathology, based on a morphology and a physiology.^ 
In his works, particularly those on suicide and the division of 
labour, he attempted to find objective criteria by which to judge 
whether a given society at a given time is normal or pathological, 
eunomic or dysnomic. For example, he tried to show that the 
increase of the rate of suicide in many countries during part of the 
nineteenth century is symptomatic of a dysnomic or, in his 
terminology, anomic, social condition. Probably there is no 
sociologist who would hold that Durkheim really succeeded in 
establishing an objective basis for a science of social pathology. ^ 

In relation to organic structures we can find strictly objective 
criteria by which to distinguish disease from health, pathological 
from normal, for disease is that which either threatens the organ- 
ism with death (the dissolution of its structure) or interferes with 
the activities which are characteristic of the organic type. Societies 
do not die in the same sense that animals die and therefore we 
cannot define dysnomia as that which leads, if unchecked, to the 
death of a society. Further, a society differs from an organism 
in that it can change its structural type, or can be absorbed as 
an integral part of a larger society. Therefore we cannot define 
dysnomia as a disturbance of the usual activities of a social type 
(as Durkheim tried to do). 

Let us return for a moment to the Greeks. They conceived 
the health of an organism and the eunomia of a society as being 
in each instance a condition of the harmonious working together 

* For what is here called dysnomia Durkheim used the term anomia (anomie 
in French). This is to my mind inappropriate. Health and disease, eunomia and 
dysnomia, are essentially relative terms. 

^ I would personally agree in the main with the criticisms of Roger Lacombe 
(La Methode Sociologique de Durkheim, 1926, ch. iv) on Durldieim's general 
theory of social pathology, and with the criticisms of Durkheim 's treatment 
of suicide presented by Halbwachs, Les Causes du Suicide. 


of its parts.^ Now this, where society is concerned, is the same 
thing as what was considered above as the functional unity or 
inner consistency of a social system, and it is suggested that for the 
degree of functional unity of a particular society it may be possible 
to establish a purely objective criterion. Admittedly this cannot be 
done at present; but the science of human society is as yet in its 
extreme infancy. So that it may be that we should say that, while 
an organism that is attacked by a virulent disease will react thereto, 
and, if its reaction fails, will die, a society that is thrown into a 
condition of functional disunity or inconsistency (for this we now 
provisionally identify with dysnomia) will not die, except in such 
comparatively rare instances as an Australian tribe oversvhelmed 
by the white man's destructive force, but will continue to struggle 
toward some sort of eunomia, some kind of social health, and may, 
in the course of this, change its structural type. This process, it 
seems, the 'functionalist' has ample opportunities of observing 
at the present day, in native peoples subjected to the domination 
of the civilised nations, and in those nations themselves. ^ 

Space will not allow a discussion here of another aspect of 
functional theory, viz. the question whether change of social type 
is or is not dependent on function, i.e. on the laws of social physio- 
logy. My own view is that there is such a dependence and that its 
nature can be studied in the development of the legal and political 
institutions, the economic systems and the religions of Europe 
through the last tw-enty-five centuries. For the preliterate societies 
with which anthropology is concerned, it is not possible to study 
the details of long processes of change of type. The one kind of 
change which the anthropologist can observe is the disintegration 
of social structures. Yet even here we can observe and compare 
spontaneous movements towards reintegration. We have, for 
instance, in Africa, in Oceania, and in America the appearance of 
new reUgions which can be interpreted on a functional hypothesis 

^ See, for example, the Fourth Book of Plato's Republic. 

^ To avoid misunderstanding it is perhaps necessary to observe that this 
distinction of eunomic and dysnomic social conditions does not give us any 
evaluation of these societies as 'good' or 'bad'. A savage tribe practising polyg- 
amy, cannibalism, and sorcery can possibly show a higher degree of functional 
unity or consistency than the United States of 1935. This objective judgment, 
for such it must be if it is to be scientific, is something very different from any 
judgment as to which of the two social systems is the better, the more to be 
desired or approved. 


as attempts to relieve a condition of social dysnomia produced 
by the rapid modification of the social Hfe through contact with 
white civilisation. 

The concept of function as defined above constitutes a 
'working hypothesis' by which a number of problems are form- 
ulated for investigation. No scientific enquiry is possible without 
some such formulation of working hypotheses. Two remarks are 
necessary here. One is that the hypothesis does not require the 
dogmatic assertion that everything in the life of every community 
has a function. It only requires the assumption that it may 
have one, and that we are justified in seeking to discover it. The 
second is that what appears to be the same social usage in two 
societies may have different functions in the two. Thus the practice 
of celibacy in the Roman Catholic Church of today has very 
different functions from those of ceUbacy in the early Christian 
Church. In other words, in order to define a social usage, and 
therefore in order to make valid comparisons between the usages 
of different peoples or periods, it is necessary to consider not 
merely the form of the usage but also its function. On this basis, 
for example, belief in a Supreme Being in a simple society is 
something different from such a belief in a modern civilised 

The acceptance of the functional hypothesis or point of view 
outlined above results in the recognition of a vast number of 
problems for the solution of which there are required wide com- 
parative studies of societies of many diverse types and also in- 
tensive studies of as many single societies as possible. In field 
studies of the simpler peoples it leads, first of all, to a direct study 
of the social life of the community as the functioning of a social 
structure, and of this there are several examples in recent literature. 
Since the function of a social activity is to be found by examining 
its effects upon individuals, these are studied, either in the average 
individual or in both average and exceptional individuals. Further, 
the hypothesis leads to attempts to investigate directly the func- 
tional consistency or unity of a social system and to determine as 
far as possible in each instance the nature of that unity. Such field 
studies will obviously be different in many ways from studies 
carried out from other points of view, e.g. the ethnological point 
of view that lays emphasis on diffusion. We do not have to say 
that one point of view is better than another, but only that they 


are different, and any particular piece of work should be judged 
in reference to what it aims to do. 

If the view here outlined is taken as one form of 'functionalism', 
a few remarks on Dr. Lesser's paper become permissible. He 
makes reference to a difference of 'content' in functional and non- 
functional anthropology. From the point of view here presented 
the 'content' or subject-matter of social anthropology is the whole 
social life of a people in all its aspects. For convenience of handling 
it is often necessary to devote special attention to some particular 
part or aspect of the social life, but if functionalism means 
anything at all it does mean the attempt to see the social life of a \ - 
people as a whole, as a functional unity. ' 

Dr. Lesser speaks of the functionalist as stressing 'the 
psychological aspects of culture', I presume that he here refers 
to the functionalist's recognition that the usages of a society work 
or 'function' only through their effects in the life, i.e. in the 
thoughts, sentiments and actions of individuals. 

The 'functionalist' point of view here presented does therefore 
imply that we have to investigate as thoroughly as possible all 
aspects of social life, considering them in relation to one another, 
and that an essential part of the task is the investigation of the 1 
individual and of the way in which he is moulded by or adjusted ^ 
to the social life. 

Turning from content to method Dr. Lesser seems to find 
some conflict between the functional point of view and the his- 
torical. This is reminiscent of the attempts formerly made to see 
a conflict between sociology and history. There need be no con- 
flict, but there is a difference. 

There is not, and cannot be, any conflict between the functional 
hypothesis and the view that any culture, any social system, is the 
end-result of a unique series of historical accidents. The process of 
development of the race-horse from its five-toed ancestor was a 
unique series of historical accidents. This does not conflict with 
the view of the physiologist that the horse of today and all the 
antecedent forms conform or conformed to physiological laws, i.e. 
to the necessary conditions of organic existence. Palaeontology and 
physiology are not in conflict. One 'explanation' of the race-horse 
is to be found in its history — how it came to be just what it is and 
where it is. Another and entirely independent 'explanation' is to 
show how the horse is a special exemplification of physiological 


laws. Similarly one 'explanation' of a social system will be its 
history, where we know it — the detailed account of how it came 
to be what it is and where it is. Another 'explanation' of the same 
system is obtained by showing (as the functionalist attempts to do) 
that it is a special exemplification of laws of social physiology 
■or social functioning. The two kinds of explanation do not 
conflict, but supplement one another.^ 

The functional hypothesis is in conflict with two views that 
are held by some ethnologists, and it is probably these, held as 
they often are without precise formulation, that are the cause of 
the antagonism to that approach. One is the ' shreds and patches' 
theory of culture, the designation being taken from a phrase of 
Professor Lowie^ when he speaks of 'that planless hodge-podge, 
that thing of shreds and patches called civilisation'. The con- 
centration of attention on what is called the diffusion of culture- 
traits tends to produce a conception of culture as a collection of 
disparate entities (the so-called traits) brought together by pure 
historical accident and having only accidental relations to one 
another. The conception is rarely formulated and maintained with 
any precision, but as a half-unconscious point of view it does 
seem to control the thinking of many ethnologists. It is, of 
course, in direct conflict with the hypothesis of the functional 
unity of social systems. 

The second view which is in direct conflict with the functional 
hypothesis is the view that there are no discoverable significant 
sociological laws such as the functionalist is seeking. I know that 

^ I see no reason at all why the two kinds of study — the historical and the 
functional — should not be carried on side by side in perfect harmony. In fact, 
for fourteen years I have been teaching both the historical and geographical 
study of peoples under the name of ethnology in close association with archae- 
ology, and the functional study of social systems under the name of social 
anthropology. I do think that there are many disadvantages in mixing the 
two subjects together and confusing them. See 'The Methods of Ethnology 
and Social Anthropology' {South African Journal of Science, 1923, pp. 124-47). 

^ Primitive Society, p. 441. A concise statement of this point of view is the 
following passage from Dr. Ruth Benedict's 'The Concept of the Guardian 
Spirit in North America' {Memoirs, American Anthropological Association, 
29, 1923), p. 84: 'It is, so far as we can see, an ultimate fact of human nature 
that man builds up his culture out of disparate elements, combining and re- 
combining them; and until we have abandoned the superstition that the result 
is an organism functionally interrelated, we shall be unable to see our cultural 
life objectively, or to control its manifestations.' I think that probably neither 
Professor Lowie nor Dr. Benedict would, at the present time, maintain this 
view of the nature of culture. 


some two or three ethnologists say that they hold this view, but 
I have found it impossible to know what they mean, or on what 
sort of evidence (rational or empirical) they would base their 
contention. Generalisations about any sort of subject matter are 
of two kinds: the generalisations of common opinion, and generali- 
sations that have been verified or demonstrated by a systematic 
examination of evidence afforded by precise observations sys- 
tematically made. Generalisations of the latter kind are called 
scientific laws. Those who hold that there are no laws of human 
society cannot hold that there are no generalisations about human 
society because they themselves hold such generalisations and 
even make new ones of their own. They must therefore hold that 
in the field of social phenomena, in contradistinction to physical 
and biological phenomena, any attempt at the systematic testing 
of existing generalisations or towards the discovery and veri- 
fication of new ones, is, for some unexplained reason, futile, or, 
as Dr. Radin puts it, 'crying for the moon'. Argument against such 
a contention is unprofitable or indeed impossible. 



IT has been suggested to me by some of my friends that I 
should use this occasion to offer some remarks about my 
own point of view in social anthropology; and since in my 
teaching, beginning at Cambridge and at the London School of 
Economics thirty years ago, I have consistently emphasised the 
importance of the study of social structure, the suggestion made to 
me was that I should say something on that subject. 

I hope you will pardon me if I begin with a note of personal 
explanation. I have been described on more than one occasion as 
belonging to something called the 'Functional School of Social 
Anthropology' and even as being its leader, or one of its leaders. 
This Functional School does not really exist; it is a myth invented 
by Professor Malinowski. He has explained how, to quote his 
own words, 'the magnificent title of the Functional School of 
Anthropology has been bestowed by myself, in a way on myself, 
and to a large extent out of my own sense of irresponsibility'. 
Professor Malinowski's irresponsibility has had unfortunate 
results, since it has spread over anthropology a dense fog of 
discussion about 'functionalism'. Professor Lowie has announced 
that the leading, though not the only, exponent of functionalism 
in the nineteenth century was Professor Franz Boas. I do not 
think that there is any sense, other than the purely chronological 
one, in which I can be said to be either the follower of Professor 
Boas or the predecessor of Professor Malinowski. The statement 
that I am a 'functionalist' would seem to me to convey no definite 

There is no place in natural science for 'schools' in this sense, 
and I regard social anthropology as a branch of natural science. 
Each scientist starts from the work of his predecessors, finds 
problems which he believes to be significant, and by obser\'ation 
and reasoning endeavours to make some contribution to a growing 
body of theory. Co-operation amongst scientists results from the 

^ Presidential Address to the Royal Anthropological Institute. Reprinted 
from the Journal of the Royal Anthropological Institute, Vol. LXX, 1940. 



fact that they are working on the same or related problems. 
Such co-operation does not result in the formation of schools, in 
the sense in which there are schools of philosophy or of painting. 
There is no place for orthodoxies and heterodoxies in science. 
Nothing is more pernicious in science than attempts to establish 
adherence to doctrines. All that a teacher can do is to assist the, 
student in learning to understand and use the scientific method. 
It is not his business to make disciples. 

I conceive of social anthropology as the theoretical natural 
science of human society, that is, the investigation of social 
phenomena by methods essentially similar to those used in 
the physical and biological sciences. I am quite willing to call the 
subject 'comparative sociolog)^', if anyone so wishes. It is the 
subject itself, and not the name, that is important. As you know, 
there are some ethnologists or anthropologists who hold that it is 
not possible, or at least not profitable, to apply to social phenomena 
the theoretical methods of natural science. For these persons 
social anthropology, as I have defined it, is something that does 
not, and never will, exist. For them, of course, my remarks will 
have no meaning, or at least not the meaning I intend them to 

While I have defined social anthropology as the study of 
human society, there are some who define it as the study of 
culture. It might perhaps be thought that this difference of 
definition is of minor importance. Actually it leads to two different 
kinds of study, between which it is hardly possible to obtain 
agreement in the formulation of problems. 

For a preliminary definition of social phenomena it seems 
sufficiently clear that what we have to deal with are relations of 
association between individual organisms. In a hive of bees 
there are the relations of association of the queen, the workers 
and the drones. There is the association of animals in a herd, of 
a mother-cat and her kittens. These are social phenomena; 
I do not suppose that anyone will call them cultural phenomena. 
In anthropology, of course, we are only concerned with human 
beings, and in social anthropology, as I define it, what we have 
to investigate are the forms of association to be found amongst 
human beings. 

Let us consider what are the concrete, observable facts with 
which the social anthropologist is concerned. If we set out to study. 


for example, the aboriginal inhabitants of a part of Australia, we 
find a certain number of individual human beings in a certain 
natural environment. We can observe the acts of behaviour of 
tfiese individuals, including, of course, their acts of speech, 
/and the material products of past actions. We do not observe a 
/ 'culture', since that word denotes, not any concrete reality, but 
an abstraction, and as it is commonly used a vague abstraction. 
But direct observation does reveal to us that these human beings 
/are connected by a complex network of social relations. I use the 
Wrm 'social structure' to denote this network of actually existing 
relations. It is this that I regard it as my business to study if I 
"Xam working, not as an ethnologist or psychologist, but as a social 
anthropologist. I do not mean that the study of social structure is 
the whole of social anthropology, but I do regard it as being in 
a very important sense the most fundamental part of the science. 
My view of natural science is that it is the systematic investiga- 
tion of the structure of the universe as it is revealed to us through 
our senses. There are certain important separate branches of science, 
each of which deals with a certain class or kind of structures, the 
aim being to discover the characteristics of all structures of that kind. 
So atomic physics deals with the structure of atoms, chemistry with 
the structure of molecules, crystallography and colloidal chemistry 
with the structure of crystals and colloids, and anatomy and 
physiology with the structures of organisms. There is, therefore, 
I suggest, place for a branch of natural science which will have 
for its task the discovery of the general characteristics of those 
social structures of which the component units are human beings. 
Social phenomena constitute a distinct class of natural 
phenomena. They are all, in one way or another, connected 
with the existence of social structures, either being implied in or 
resulting from them. Social structures are just as real as are 
individual organisms. A complex organism is a collection of 
living cells and interstitial fluids arranged in a certain structure; 
and a living cell is similarly a structural arrangement of complex 
molecules. The physiological and psychological phenomena that 
we observe in the lives of organisms are not simply the result of the 
nature of the constituent molecules or atoms of which the organism 
is built up, but are the result of the structure in which they are 
united. So also the social phenomena which we observe in any 
human society are not the immediate result of the nature of 


individual human beings, but are tlie result of the social structure 
by which they are united. 

It should be noted that to say we are studying social structures 
is not exactly the same thing as saying that we study social relations, 
which is how some sociologists define their subject. A particular 
social relation between two persons (unless they be Adam and 
Eve in the Garden of Eden) exists only as part of a wide network 
of social relations, involving many other persons, and it is this 
network which I regard as the object of our investigations. 

I am aware, of course, that the term 'social structure' is used 
in a number of different senses, some of them very vague. This 
is unfortunately true of many otlier terms commonly used by 
anthropologists. The choice of terms and their definitions is a 
matter of scientific convenience, but one of the characteristics of a 
science as soon as it has passed the first formative period is the 
existence of technical terms which are used in the same precise 
meaning by all the students of that science. By this test, I regret 
to say, social anthropology reveals itself as not yet a formed 
science. One has therefore to select for oneself, for certain terms, 
definitions which seem to be the most convenient for the purpose 
of scientific analysis. 

There are some anthropologists who use the term social 
structure to refer only to persistent social groups, such as nations, 
tribes and clans, which retain their continuity, their identity 
as individual groups, in spite of changes in their membership. 
Dr. Evans-Pritchard, in his recent admirable book on the Nuer, 
prefers to use the term social structure in this sense. Certainly 
the existence of such persistent social groups is an exceedingly 
important aspect of structure. But I find it more useful to include 
under the term social structure a good deal more than this. 

In the first place, I regard as a part of the social structure all 
social relations of person to person. For example, the kinship 
structure of any society consists of a number of such dyadic 
relations, as between a father and son, or a mother's brother and 
his sister's son. In an Australian tribe the whole social structure 
is based on a network of such relations of person to person, 
established through genealogical connections. 

Secondly, I include under social structure the differentiation 
of individuals and of classes by their social role. The differential 
social positions of men and women, of chiefs and commoners, 


of employers and employees, are just as much determinants of 
social relations as belonging to different clans or different nations. 

In the study of social structure the concrete reality with which 
we are concerned is the set of actually existing relations, at a given 
moment of time, which link together certain human beings. 
It is on this that we can make direct observations. But it is not 
this that we attempt to describe in its particularity. Science 
(as distinguished from history or biography) is not concerned 
with the particular, the unique, but only with the general, with 
kinds, with events which recur. The actual relations of Tom, 
Dick and Harry or the behaviour of Jack and Jill may go down in 
our field note-books and may provide illustrations for a general 
description. But what we need for scientific purposes is an account 
of the form of the structure. For example, if in an Australian 
tribe I observe in a number of instances the behaviour towards 
one another of persons who stand in the relation of mother's 
brother and sister's son, it is in order that I may be able to 
record as precisely as possible the general or normal form of this 
relationship, abstracted from- the variations of particular instances, 
though taking account of those variations. 

This important distinction, between structure as an actually 
existing concrete reality, to be directly observed, and structural 
form, as what the field-worker describes, may be made clearer 
perhaps by a consideration of the continuity of social structure 
through time, a continuity which is not static like that of a hjiilding, 
but a dynamic continuity, like that of the organic structure of a 
living body. Throughout the life of an organism its structure is 
being constantly renewed; and similarly the social life constantly 
renews the social structure. Thus the actual relations of persons 
and groups of persons change from year to year, or even from day 
to day. New members come into a community by birth or im- 
migration; others go out of it by death or emigration. There are 
marriages and divorces. Friends may become enemies, or enemies 
may make peace and become friends. But while the actual structure 
changes in this way, the general structural form may remain 
relatively constant over a longer or shorter period of time. Thus 
if I visit a relatively stable community and revisit it after an interval 
of ten years, I shall find that many of its members have died and 
others have been born; the members who still survive are now ten 
years older and their relations to one another may have changed 


in many ways. Yet I may find that the kinds of relations that I 
can observe are very little different from those observed ten years 
before. The structural form has changed little. 

But, on the other hand, the structural form may change, 
sometimes gradually, sometimes with relative suddenness, as in 
revolutions and military conquests. But even in the most revolu- 
tionary changes some continuity of structure is maintained. 

I must say a few words about the spatial aspect of social 
structure. It is rarely that we find a community that is absolutely 
isolated, having no outside contact. At the present moment of 
history, the network of social relations spreads over the whole 
world, without any absolute solution of continuity anywhere. 
This gives rise to a difficulty which I do not think that sociologists 
have really faced, the difficulty of defining what is meant by the 
term 'a society'. They do commonly talk of societies as if they 
were distinguishable, discrete entities, as, for example, when we 
are told that a society is an organism. Is the British Empire a 
society or a collection of societies? Is a Chinese village a society, 
or is it merely a fragment of the Republic of China? 

If we say that our subject is the study and comparison of 
human societies, we ought to be able to say what are the unit 
entities with which we are concerned. 

If we take any convenient locality of a suitable size, we can 
study the structural system as it appears in and from that region, 
i.e. the network of relations connecting the inhabitants amongst 
themselves and with the people of other regions. We can thus 
observe, describe, and compare the systems of social structure of 
as many localities as we wish. To illustrate what I mean, I may 
refer to two recent studies from the University of Chicago, one 
of a Japanese village, Suye Mura, by Dr. John Embree, and the 
other of a French Canadian community, St. Denis, by Dr. 
Horace Miner. 

Closely connected with this conception of social structure 
is the conception of 'social personality' as the position occupied 
by a human being in a social structure, the complex formed by 
all his social relations with others. Every human being living in 
society is two things: he is an individual and also a person. As an 
individual, he is a biological organism, a collection of a vast number 
of molecules organised in a complex structure, within which, 
as long as it persists, there occur physiological and psychological 



actions and reactions, processes and changes. Human beings 
as individuals are objects of study for physiologists and psycho- 
logists. The human being as a person is a complex of social 
relationships. He is a citizen of England, a husband and a father, 
a bricklayer, a member of a particular Methodist congregation, 
a voter in a certain constituency, a member of his trade union, an 
adherent of the Labour Party, and so on. Note that each of these 
descriptions refers to a social relationship, or to a place in a social 
structure. Note also that a social personality is something that 
changes during the course of the life of the person. As a person, 
the human being is the object of study for the social anthropologist. 
We cannot study persons except in terms of social structure, nor 
can we study social structure except in terms of the persons who 
are the units of which it is composed. 

If you tell me that an individual and a person are after all 
really the same thing, I would remind you of the Christian creed. 
God is three persons, but to say that He is three individuals 
is to be guilty of a heresy for which men have been put to death. 
Yet the failure to distinguish individual and person is not merely 
a heresy in religion; it is worse than that; it is a source of confusion 
in science. 

I have now sufficiently defined, I hope, the subject-matter 
of what I regard as an extremely important branch of social 
anthropology. The method to be adopted follows immediately 
from this definition. It must combine with the intensive study of 
single societies (i.e. of the structural systems observable in par- 
ticular communities) the systematic comparison of many societies 
(or structural systems of different types). The use of comparison 
is indispensable. The study of a single society may provide 
materials for comparative study, or it may afford occasion for 
hypotheses, which then need to be tested by reference to other 
societies; it cannot give demonstrated results. 

Our first task, of course, is to learn as much as we can about 
the varieties, or diversities, of structural systems. This requires 
field research. Many writers of ethnographical descriptions do not 
attempt to give us any systematic account of the social structure. 
But a few social anthropologists, here and in America, do recognise 
the importance of such data and their work is providing us with a 
steadily growing body of material for our study. Moreover, their 
researches are no longer confined to what are called 'primitive' 


societies, but extend to communities in such regions as Sicily, 
Ireland, Japan, Canada and the United States. 

If we are to have a real comparative morphology of societies, 
however, we must aim at building up some sort of classification of 
types of structural systems. That is a complex and difficult task, 
to which I have myself devoted attention for thirty years. It is the 
kind of task that needs the co-operation of a number of students 
and I think I can number on my fingers those who are actively 
interested in it at the present time. Nevertheless, I believe some 
progress is being made. Such work, however, does not produce 
spectacular results and a book on the subject would certainly 
not be an anthropological best-seller. 

We should remember that chemistry and biology did not 
become fully formed sciences until considerable progress had been 
made with the systematic classification of the things they were 
dealing with, substances in the one instance and plants and animals 
in the other. 

Besides this morphological study, consisting in the definition, 
comparison and classification of diverse structural systems, there 
is a physiological study. The problem here is: How do structural 
systems persist? What are the mechanisms which maintain a 
network of social relations in existence, and how do they work? 
In using the terms morphology and physiology, I may seem to be 
returning to the analogy between society and organism which was 
so popular with medieval philosophers, was taken over and often 
misused by nineteenth century sociologists, and is completely 
rejected by many modern writers. But analogies, properly used, 
are important aids to scientific thinking and there is a real and 
significant analogy between organic structure and social structure. 

In what I am thus calling social physiology we are concerned 
not only with social structure, but with every kind of social 
phenomenon. Morals, law, etiquette, religion, government, and 
education are all parts of the complex mechanism by which a 
social structure exists and persists. If we take up the structural 
point of view, we study these things, not in abstraction or isolation, 
but in their direct and indirect relations to social structure, i.e. 
with reference to the way in which they depend upon, or affect, 
the social relations between persons and groups of persons. 
I cannot do more here than offer a few brief illustrations of \\ hat 
this means. 


Let US first consider the study of language. A language is a 
connected set of speech usages observed within a defined speech- 
community. The existence of speech-communities and their 
sizes are features of social structure. There is, therefore, a certain 
very general relation between social structure and language. But 
if we consider the special characteristics of a particular language — 
its phonology, its morphology and even to a great extent its 
vocabulary — there is no direct connection of either one-sided 
or mutual determination between these and the special charac- 
teristics of the social structure of the community within which the 
language is spoken. We can easily conceive that two societies 
might have very similar forms of social structure and very different 
kinds of language, or vice versa. The coincidence of a particular 
form of social structure and a particular language in a given 
community is always the result of historical accident. There may, 
of course, be certain indirect, remote interactions between social 
structure and language, but these would seem to be of minor 
importance. Thus the general comparative study of languages 
can be profitably carried out as a relatively independent branch 
of science, in which the language is considered in abstraction 
from the social structure of the community in which it is spoken. 

But, on the other hand, there are certain features of linguistic 
history which are specifically connected with social structure. 
As structural phenomena may be instanced the process by which 
Latin, from being the language of the small region of Latium, 
became the language of a considerable part of Europe, displacing 
the other Italic languages, Etruscan, and many Celtic languages; 
and the subsequent reverse process by which Latin split up into 
a number of diverse local forms of speech, which ultimately 
became the various Romance languages of today. 

Thus the spread of language, the unification of a number of 
separate communities into a single speech-community, and the 
reverse process of subdivision into different speech-communities, 
are phenomena of social structure. So also are those instances in 
which, in societies having a class structure, there are differences 
.of speech usage in different classes. 

I have considered language first, because linguistics is, I 
think, the branch of social anthropology which can be most 
profitably studied without reference to social structure. There is a 
reason for this. The set of speech usages which constitute a 


language does form a system, and systems of this kind can be 
compared in order to discover their common general, or abstract, 
characters, the determination of which can give us laws, which will 
be specifically laws of linguistics. 

Let us consider very briefly certain other branches of social 
anthropology and their relation to the study of social structure. 
If we take the social life of a local community over a period, let us 
say a year, we can observe a certain sum total of activities carried 
out by the persons who compose it. We can also observe a certain 
apportionment of these activities, one person doing certain things, 
another doing others. This apportionment of activities, equivalent 
to what is sometimes called the social division of labour, is an 
important feature of the social structure. Now activities are carried 
out because they provide some sort of ' gratification ', as I propose 
to call it, and the characteristic feature of social life is that 
activities of certain persons provide gratifications for other 
persons. In a simple instance, when an Australian blackfcllow 
goes hunting, he provides meat, not only for himself, but for 
his wife and children and also for other relatives to whom it is 
his duty to give meat when he has it. Thus in any society there 
is not only an apportionment of activities, but also an apportion- 
ment of the gratifications resulting therefrom, and some sort of 
social machinery, relatively simple or, sometimes, highly complex, 
by which the system works. 

It is this machinery, or certain aspects of it, that constitutes 
the special subject-matter studied by the economists. They con- 
cern themselves with what kinds and quantities of goods are 
produced, how they are distributed (i.e. their flow from person 
to person, or region to region), and the way in which they are 
disposed oL Thus what are called economic institutions are 
extensively studied in more or less complete abstraction from the 
rest of the social system. This method does undoubtedly provide 
useful results, particularly in the study of complex modern 
societies. Its weaknesses become apparent as soon as we attempt 
to apply it to the exchange of goods in what are called primitive 

The economic machinery of a society appears in quite a new 
light if it is studied in relation to the social structure. The exchange 
of goods and services is dependent upon, is the result of, and at 
the same time is a means of maintaining a certain structure, a 


network of relations between persons and collections of persons. 
For the economists and politicians of Canada the potlatch of the 
Indians of the north-west of America was simply wasteful foolish- 
ness and it was therefore forbidden. For the anthropologist it 
was the machinery for maintaining a social structure of lineages, 
clans and moieties, with which was combined an arrangement of 
rank defined by privileges. 

Any full understanding of the economic institutions of human 
societies requires that they should be studied from two angles. 
From one of these the economic system is viewed as the mechanism 
by which goods of various kinds and in various quantities are 
produced, transported and transferred, and utilised. From the 
other the economic system is a set of relations between persons and 
groups which maintains^ and is maintained by, this exchange or 
circulation of goods and services. From the latter point of view, 
the study of the economic life of societies takes its place as part 
of the general study of social structure. 

Social relations are only observed, and can only be described, 
by reference to the reciprocal behaviour of the persons related. 
The form of a social structure has therefore to be described by 
the patterns of behaviour to which individuals and groups con- 
form in their dealings with one another. These patterns are 
partially formulated in rules which, in our own society, we dis- 
tinguish as rules of etiquette, of morals and of law. Rules, of course, 
only exist in their recognition by the members of the society; 
either in their verbal recognition, when they are stated as rules, 
or in their observance in behaviour. These two modes of re- 
cognition, as every field-worker knows, are not the same thing and 
both have to be taken into account. 

If I say that in any society the rules of etiquette, morals and 
law are part of the mechanism by which a certain set of social 
relations is maintained in existence, this statement will, I suppose, 
be greeted as a truism. But it is one of those truisms which many 
writers on human society verbally accept and yet ignore in 
theoretical discussions, or in their descriptive analyses. The 
point is not that rules exist in every society, but that what we need 
to know for a scientific understanding is just how these things 
work in general and in particular instances. 

Let us consider, for example, the study of law. If you examine 
the literature on jurisprudence you will find that legal institutions 


are studied for the most part in more or less complete abstraction 
from the rest of the social system of which they are a part. This 
is doubtless the most convenient method for la\\'yers in their 
professional studies. But for any scientific investigation of the 

• nature of law it is insufficient. The data with which a scientist 
must deal are events which occur and can be observed. In the 
field of law, the events which the social scientist can observe and 
thus take as his data are the proceedings that take place in courts 
of justice. These are the reality, and for the social anthropologist 
they are the mechanism or process by which certain definable 
social relations between persons and groups are restored, main- 
tained or modified. Law is a part of the machinery by which a ' 
certain social structure is maintained. The system of laws of a 
particular society can only be fully understood if it is studied j 
in relation to the social structure, and inversely the understanding 
of the social structure requires, amongst other things, a systematic I 
study of the legal institutions. — ^ 

I have talked about social relations, but I have not so far 
offered you a precise definition. A social relation exists between 
two or more individual organisms when there is some adjustment 
of their respective interests, by convergence of interest, or by 

\, limitation of conflicts that might arise from divergence of interests. 
I use the term 'interest' here in the widest possible sense, to 
refer to all behaviour that we regard as purposive. To speak of an 
interest implies a subject and an object and a relation between 
them. Whenever we say that a subject has a certain interest in an 

, object we can state the same thing by saying that the object has 
a certain value for the subject. Interest and value are correlative "] 
terms, which refer to the two sides of an asymmetrical relation. -J 

Thus the study of social structure leads immediately to the 
study of interests or values as the determinants of social relations. 
A social relation does not result from similarity of interests, but^^ 
rests either on the mutual interest of persons in one another, or 
on one or more common interests, or on a combination of both 
of these. The simplest form of social solidarity is where two 
persons are both interested in bringing about a certain result and 
co-operate to that end. When two or more persons have a common— y 
interest in an object, that object can be said to have a social ralue 
for the persons thus associated. If, then, practically all the ^ 
members of a society have an interest in the obscr\'ance of the 


I ^laws, we can say that the law has a social value. The study of 
social values in this sense is therefore a part of the study of social 
^ structure. 

It was from this point of view that in an early work I approached 
the study of what can conveniently be called ritual values, i.e. the 
values expressed in rites and myths. It is perhaps again a truism 
to say that religion is the cement which holds society together. 
But for a scientific understanding we need to know just how it 
does this, and that is a subject for lengthy investigations in many 
different forms of society. 

As a last example let me mention the study of magic and 
witchcraft, on which there is an extensive anthropological 
literature. I would point to Dr. Evans-Pritchard's work on the 
Zande as an illuminating example of what can be done when these 
things are systematically investigated in terms of the part they 
play in the social relations of the members of a community. 

From the point of view that I have attempted briefly to de- 
scribe, social institutions, in the sense of standardised modes of 
behaviour, constitute the machinery by which a social structure, 
a network of social relations, maintains its existence and its 
continuity. I hesitate to use the term 'function', which in recent 
years has been so much used and misused in a multitude of 
meanings, many of them very vague. Instead of being used, as 
scientific terms ought to be, to assist in making distinctions, it is 
now used to confuse things that ought to be distinguished. For it 
is often employed in place of the more ordinary words 'use' 
'purpose', and 'meaning'. It seems to me more convenient and 
sensible, as well as more scholarly, to speak of the use or uses 
of an axe or digging stick, the meaning of a word or symbol, the 
purpose of an act of legislation, rather than to use the word 
function for these various things. 'Function' has been a very 
useful technical term in physiology and by analogy with its use 
in that science it would be a very convenient means of expressing 
an important concept in social science. As I have been accustomed 
to use the word, following Durkheim and others, I would define 
- the social function of a socially standardised mode of activity, or 
f mode of thought, as its relation to the social structure to the 
existence and continuity of which it makes some contribution. 
Analogously, in a living organism, the physiological function of 
the beating of the heart, or the secretion of gastric juices, is its 


relation to the organic structure to the existence or continuity of 
which it makes its contribution. It is in this sense that I am inter- 
ested in such things as the social function of the punishment of 
crime, or the social function of the totemic rites of Australian 
tribes, or of the funeral rites of the Andaman Islanders. But 
this is not what either Professor Malinowski or Professor Lowie 
means by functional anthropology. 

Besides these two divisions of the study of social structure, 
which I have called social morphology and social physiology, 
there is a third, the investigation of the processes by which social 
structures change, of how new forms of structures come into 
existence. Studies of social change in the non-literate societies 
have necessarily been almost entirely confined to one special 
kind of process of change, the modification of the social life under 
the influence or domination of European invaders or conquerors. 

It has recently become the fashion amongst some anthropo- 
logists to treat changes of this kind in terms of what is called 
'culture contact'. By that term we can understand the one-sided 
or two-sided effects of interaction between two societies, groups, 
classes or regions having different forms of social life, different 
institutions, usages and ideas. Thus in the eighteenth century 
there was an important exchange of ideas between France and 
Great Britain, and in the nineteenth century there was a marked 
influence of German thought on both France and England. 
Such interactions are, of course, a constant feature of social life, 
but they need not necessarily involve any marked change of 
social structure. 

The changes that are taking place in the non-literate peoples 
of Africa are of a very different kind. Let us consider an African 
colony or possession of a European nation. There is a region 
that was formerly inhabited by Africans with their own social struc- 
ture. Europeans, by peaceful or forceful means, establish control 
over the region, under what we call a 'colonial' regime. A new 
social structure comes into existence and then undergoes develop- 
ment. The population now includes a certain number of Europeans 
— government officials, missionaries, traders and in some in- 
stances settlers. The social life of the region is no longer simply 
a process depending on the relations and interactions of the 
native peoples. There grows up a new political and economic 
structure in which the Europeans, even though few in numbers, 


exercise dominating influence. Europeans and Africans constitute 
different classes within the new structure, with different languages, 
different customs and modes of life, and different sets of ideas and 
values. A convenient term for societies of this kind would be 
'composite' societies; the term 'plural' societies has also been 
suggested. A complex example of a composite society is provided 
by the Union of South Africa w4th its single political and eco- 
nomic structure and a population including English-speaking 
and Afrikaans-speaking peoples of European descent, the so-called 
'coloured people' of the Cape Province, progeny of Dutch and 
Hottentots, the remaining Hottentots, the 'Malays' of Cape 
Town, descendants of persons from the Malay Archipelago, Hindus 
and Mohammedans from India and their descendants, and a 
number of Bantu tribes who constitute the majority of the 
population of the Union taken as a whole. 

The study of composite societies, the description and analysis 
of the processes of change in them, is a complex and difficult 
task. The attempt to simplify it by considering the process as 
being one in which two or more 'cultures' interact, which is the 
method suggested by Malinowski in his Introduction to Memoran- 
dum XV of the International Institute of African Language 
and Culture on 'Methods of Study of Culture Contact in Africa' 
(1938), is simply a way of avoiding the reality. For what is 
happening in South Africa, for example, is not the interaction 
of British culture, Afrikander (or Boer) culture, Hottentot cultures- 
various Bantu cultures and Indian culture, but the interaction 
of individuals and groups within an established social structure 
which is itself in process of change. What is happening in a 
Transkeian tribe, for example, can only be described by recognising 
that the tribe has been incorporated into a wide political and 
economic structural system. 

For the scientific study of primitive societies in conditions in 
which they are free from the domination by more advanced 
societies which result in these composite societies, we have 
unfortunately an almost complete lack of authentic historical 
data. We cannot study, but can only speculate about, the processes 
of change that took place in the past of which we have no record. 
Anthropologists speculate about former changes in the societies 
of the Australian aborigines, or the inhabitants of Melanesia, 
but such speculations are not history and can be of no use in 


science. For the study of social change in societies other than the 
composite societies to which reference has been made we have 
to rely on the work of historians dealing with authentic records. 

You are aware that in certain anthropological circles the term 
'evolutionary anthropologist' is almost a term of abuse. It is 
applied, however, without much discrimination. Thus Lewis 
Morgan is called an evolutionist, although he rejected the theory 
of organic evolution and in relation to society believed, not in 
evolution, but in progress, which he conceived as the steady 
material and moral improvement of mankind from crude stone 
implements and sexual promiscuity to the steam engines and 
monogamous marriage of Rochester, N.Y. But even such anti- 
evolutionists as Boas believe in progress. 

It is convenient, I think, to use the term 'progress' for the 
process by which human beings attain to greater control over the 
physical environment through the increase of knowledge and 
improvement of technique by inventions and discoveries. Th 
way in which we are now able to destroy considerable portions o 
cities from the air is one of the latest striking results of progress 
Progress is not the same thing as social evolution, but it is 
very closely connected with it. 

Evolution, as I understand the term, refers specifically to a 
process of emergence of new forms of structure. Organic 
evolution has two important features: (i) in the course of it a 
small number of kinds of organisms have given rise to a very much 
larger number of kinds; (2) more complex forms of organic 
structure have come into existence by development out of simpler 
forms. While I am unable to attach any definite meaning to such 
phrases as the evolution of culture or the evolution of language, 
I think that social evolution is a reality which the social anthropo- 
logist should recognise and study. Like organic evolution, it can 
be defined by two features. There has been a process by which, 
from a small number of forms of social structure, many different 
forms have arisen in the course of history; that is, there has been a 
process of diversification. Secondly, throughout this process more 
complex forms of social structures have developed out of, or 
replaced, simpler forms. 

Just how structural systems are to be classified with reference 
to their greater or less complexity is a problem requiring in- 
vestigation. But there is evidence of a fairly close correlation 


between complexity and anotlitr feature of structural systems, 
namely, the extent of the field of social relations. In a structural 
system with a narrow total social field, an average or typical 
person is brought into direct and indirect social relations with 
only a small number of other persons. In systems of this type we 
may find that the linguistic community — the body of persons 
who speak one language — numbers from 250 to 5C0, while the 
political community is even smaller, and economic relations by the 
exchange of goods and services extend only over a very narrow 
range. Apart from the differentiation by sex and age, there is very 
little differentiation of social role between persons or classes. We 
can contrast with this the systems of social structure that we observe 
today in England or the United States. Thus the process of human 
history to which I think the term social evolution may be ap- 
propriately applied might be defined as the process by which 
wide-range systems of social structure have grown out of, or 
replaced, narrow-range systems. Whether this view is acceptable 
or not, I suggest that the concept of social evolution is one which 
requires to be defined in terms of social structure. 

There is no time on this occasion to discuss the relation of the 
study of social structure to the study of culture. For an interesting 
attempt to bring the two kinds of study together I would refer 
you to Mr. Gregory Bateson's book N avert. I have made no attempt 
to deal with social anthropology as a whole and with all its various 
branches and divisions. I have endeavoured only to give you a very 
general idea of the kind of study to which I have found it scienti- 
fically profitable to devote a considerable and steadily increasing 
proportion of my time and energy. The only reward that I have 
sought I think I have in some measure found — something of the 
kind of insight into the nature of the world of which we are part 
that only the patient pursuit of the method of natural science can 



IN any community there are certain modes of behaviour which 
are usual and which characterise that particular community. 
Such modes of behaviour may be called usages. All social 
usages have behind them the authority of the society, but among 
them some are sanctioned and others are not. A sanction is a 
reaction on the part of a society or of a considerable number of its 
members to a mode of behaviour which is thereby approved 
(positive sanctions) or disapproved (negative sanctions). Sanctions 
may further be distinguished according to whether they are 
diffuse or organised; the former are spontaneous expressions of 
approval or disapproval by members of the community acting 
as individuals, while the latter are social actions carried out 
according to some traditional and recognised procedure. It is a 
significant fact that in all human societies the negative sanctions 
are more definite than the positive. Social obligations may be 
defined as rules of behaviour the failure to observe which entails 
a negative sanction of some sort. These are thus distinguished 
from non-obligatory social usages, as, for example, customary 
technical procedures. 

The sanctions existing in a community constitute motives in 
the individual for the regulation of his conduct in conformity with 
usage. They are effective, first, through the desire of the individual 
to obtain the approbation and to avoid the disapprobation of his 
fellows, to win such rewards or to avoid such punishment as the 
community offers or threatens; and, second, through the fact 
that the individual learns to react to particular modes of behaviour 
with judgments of approval and disapproval in the same way as do 
his fellows, and therefore measures his own behaviour both in 
anticipation and in retrospect by standards which conform more 
or less closely to those prevalent in the community to which he 
belongs. What is called conscience is thus in the widest sense the 
reflex in the individual of the sanctions of the society. 

^ Reprinted from the Encyclopeedia of the Social Sciences, Macmillan Co. , 
New York, 1933, Vol. XIII, pp. 531-4. 



It is convenient to begin a discussion of sanctions by a 
consideration of the diffuse negative sanctions, comprising 
reactions tovi^ard the particular or general behaviour of a member 
of the community which constitute judgments of disapproval. 
In such reactions there are not only differences of degree — for 
disapproval is felt and expressed with different degrees of in- 
tensity — but also differences of kind. Such differences are difficult 
to define and classify. In the English language, for example, 
there are a large number of words which express disapproval of 
individual behaviour; these vary from discourteous, unmannerly, 
unseemly and unworthy, through improper, discreditable, 
dishonourable and disreputable, to outrageous and infamous. 
Every society or culture has its own ways of judging behaviour, 
and these might conveniently be studied in the first instance 
through the vocabulary. But until comparative study of societies 
of different types has proceeded further no systematic classification 
of the kinds of diffuse negative sanction is possible. Provisionally 
the negative moral or ethical sanction may be defined as a reaction 
of reprobation by the community toward a person whose conduct 
is disapproved; moral obligations may thus be considered as 
rules of conduct which, if not observed, bring about a reaction 
of this kind. Another distinguishable sanction is that whereby 
the behaviour of an individual is met with ridicule on the 
part of his fellows; this has been called the satirical sanction. 
The varieties of diffuse positive sanctions, being less definite 
than negative sanctions, are therefore still more difficult to 

From the diffuse sanctions already described there should be 
distinguished what may be called (by a wide extension of the 
term) religious sanctions; these have also been named supernatural 
sanctions and mystic sanctions, but both these terms have 
unsatisfactory connotations. The religious sanctions are constituted 
in any community by the existence of certain beliefs which are 
themselves obligatory; it is therefore only within a religious 
community that these sanctions exist. They take the form that 
certain deeds by an individual produce a modification in his 
religious condition, in either a desirable (good) or an undesirable 
(evil) direction. Certain acts are regarded as pleasing to gods or 
spirits or as establishing desirable relations with them, while 
others displease them or destroy in some way the desirable 


harmonious relations. The reUgious condition of the individual 
is in these instances conceived to be determined by his relation 
to personal spiritual beings. The change in the religious condition 
may elsewhere be regarded as the immediate effect of the act itself, 
not mediated by its effects on some personal god or spirit, a view 
common not only in many of the simpler societies, but also found 
in a special form in Buddhism and in other advanced Indian 
religions. Sin may be defined as any mode of behaviour which falls 
under a negative religious sanction; there is no convenient term 
for the opposite of sin, that is an action which produces religious 
merit or a desirable ritual condition. 

The religious sanctions involve the belief that most un- 
satisfactory ritual or religious conditions (pollution, uncleanness, 
sinfulness) can be removed or neutralised by socially prescribed 
or recognised procedures, such as lustration, sacrifice, penance, 
confession and repentance. These expiatory rites are also con- 
sidered to act either immediately, or mediately through their 
effects on gods or spirits, depending upon whether the sin is 
regarded as acting in the one way or the other. 

While in modem western civilisation a sin is usually regarded 
as necessarily a voluntary action or thought, in many simple 
societies an involuntary action may fall within the given definition 
of sin. Sickness — for example, leprosy among the Hebrews — is 
often regarded as similar to ritual or religious pollution and as 
therefore requiring expiation or ritual purification. A condition of 
ritual or rehgious impurity is normally considered as one of im- 
mediate or ultimate danger to the individual; it may be believed 
that he will fall sick and perhaps die unless he can be purified. 
In some religions the religious sanction takes the form of a belief 
that an individual who sins in this life will suffer some form of 
retribution in an after-life. In many instances an individual 
who is ritually unclean is looked upon as a source of danger not 
only to himself but also to those with whom he comes in contact 
or to the whole community. He may therefore be more or less 
excluded for a time or even permanently from participation in the 
social life of the community. Frequently, if not always, an 
obligation therefore rests upon the sinner, or unclean person, 
to undertake the necessary process of purification. 

Thus the religious sanctions differ from the other diffuse 
sanctions by reason of the beliefs and conceptions indicated above, 


which cannot be defined or described in any simple way. Somewhat 
similar beliefs underlie magical practices and procedure in relation 
to luck, but whereas religious observances and the beliefs associated 
with them are obligatory within a given religious community, 
the former are comparable with technical procedures, customary 
but not obligatory. 

Organised sanctions are to be regarded as special developments 
of the diffuse sanctions, frequently under the influence of the 
beliefs belonging to religion. Organised positive sanctions, or 
premial sanctions, are rarely developed to any great extent. 
Honours, decorations, titles and other rewards for merit, including 
monetary rewards such as special pensions, given to individuals 
by a community as a whole, are characteristic of modem societies. 
In preliterate societies a man who has slain an enemy may be 
given the right to distinguish himself by wearing some special 
decoration or in other ways. 

Organised negative sanctions, important among which are the 
penal sanctions of criminal law, are definite recognised procedures 
directed against persons whose behaviour is subject to social 
disapproval. There are many varieties of such procedures, the most 
important and widespread being the following: subjection to 
open expression of reprobation or derision, as, for example, 
through forcible public exposure by confinement in stocks; 
partial exclusion, permanent or temporary, from full participation 
in social life and its privileges, including permanent or temporary 
loss of civil or religious rights; specific loss of social rank, or 
degradation, the exact contrary of the positive sanction of pro- 
motion; infliction of loss of property by imposition of a fine or by 
forcible seizure or destruction; infliction of bodily pain; mutilation 
or branding in which pain is incidental to permanent exposure to 
reprobation; permanent exclusion from the community, as by 
exile; imprisonment; and punishment by death. These sanctions 
are legal sanctions when they are imposed by a constituted 
authority, political, military or ecclesiastic. 

In any given society the various primary sanctions form a more 
or less systematic whole which constitutes the machinery of 
social control. There is an intimate relation between the religious 
sanctions and the moral sanctions, which varies, however, in 
diff"erent societies, and cannot be stated in any brief formula. 
The primary legal sanctions of criminal law, in all societies 


except the highly secularised modern states, show a close con- 
nection with religious beliefs. 

Besides these primary social sanctions and resting upon them 
there are certain sanctions which may be termed secondary; 
these are concerned with the actions of persons or groups in their 
effects upon other persons or groups. In modern civil law, for 
example, when an individual is ordered by a court to pay damages, 
the primary sanction behind the order is the power of the court 
to make forcible seizure of his property or to imprison or otherwise 
punish him for contempt of court if he fails to obey. Thus 
secondary sanctions consist of procedures carried out by a 
community, generally through its representatives, or by individuals 
with the approval of the community, when recognised rights 
have been infringed. They are based upon the general principle 
that any person who has suffered injury is entitled to satisfaction 
and that such satisfaction should be in some way proportioned to 
the extent of the injury. 

One class of such procedures consists of acts of retaliation, by 
which is meant socially approved, controlled and limited acts of 
revenge. Thus in an Australian tribe when one man has committed 
an offence against another, the latter is permitted by public 
opinion, often definitely expressed by the older men, to throw a 
certain number of spears or boomerangs at the former or in some 
instances to spear him in the thigh. After he has been given such 
satisfaction he may no longer harbour ill feelings against the 
offender. In many preliterate societies the killing of an individual 
entitles the group to which he belongs to obtain satisfaction by 
killing the offender or some member of his group. In regulated 
vengeance the offending group must submit to this as an act of 
justice and must not attempt further retaliation. Those who have 
received such satisfaction are felt to have no further grounds for 
ill feeling. 

Satisfaction for injury may be obtained also through the duel, a 
recognised and controlled combat between individuals, or through 
similar combats between two groups. Among Australian tribes 
duelling with spears, boomerangs, clubs and shields or stone 
knives, with the bystanders ready to interfere if they think things 
are going too far, is a frequently adopted alternative to one-sided 
retaliation. In these same tribes there are similar regulated 
combats between two groups, sometimes in the presence of other 


groups who see that there is fair play. It is often difficult to draw a 
dividing line between such group combats and warfare; in fact 
they may possibly be regarded as a special kind of warfare 
characteristic of primitive rather than of civilised societies. 
Frequently, therefore, war may be regarded as a secondary 
social sanction similar to the duel. A political group maintains 
recognition of its rights by the threat of war if those rights should 
be infringed. Even in the simplest societies it is recognised that 
certain acts are right in war and others are wrong and that a 
declaration of war may be just in certain circumstances and in 
others unjust, so that the conduct of warfare is to some extent 
controlled by diffuse sanctions. 

Indemnification is often found as an alternative to retaliation as 
a means of giving and receiving satisfaction. An indemnity is 
something of value given by a person or group to another person or 
group in order to remove or neutralise the effects of an infringe- 
ment of rights. It may be distinguished from a propitiatory gift 
by the fact that it is obligatory (i.e. subject to a negative sanction, 
diffuse or organised) in the particular circumstances. A payment 
made in anticipation of an invasion of rights with the consent 
of the person or persons receiving it may be regarded as an in- 
demnity. Thus in many societies taking a woman in marriage 
is regarded as an invasion of the rights of her family and kin, 
so that before they consent to part with her they must receive 
an indemnity or the promise of such. In these cases the process of 
indemnification bears some similarity to that of purchase, which is 
a transfer of rights of property for a consideration. 

In many preliterate societies procedures of indemnification 
are carried out under the diffuse sanction of public opinion, 
which compels an individual to indemnify one whose rights he has 
infringed. In some societies there is a recognised right of an injured 
person to indemnify himself by forcible seizure of the property of 
the offender. When society becomes politically organised, pro- 
cedures of retaliation and indemnification, backed by diffuse 
sanctions, give place to legal sanctions backed by the power of 
judicial authorities to inflict punishment. Thus arises civil law, 
by which a person who has suffered an infringement of rights 
may obtain reparation or restitution from the person responsible. 

In a consideration of the functions of social sanctions it is not 
the effects of the sanction upon the person to whom they are 


applied that are most important but rather the general effects 
within the community applying the sanctions. For the application 
of any sanction is a direct affirmation of social sentiments by 
the community and thereby constitutes an important, possibly 
essential, mechanism for maintaining these sentiments. Organised 
negative sanctions in particular, and to a great extent the secondary 
sanctions, are expressions of a condition of social dysphoria 
brought about by some deed. The function of the sanction is to 
restore the social euphoria by giving definite collective expression 
to the sentiments which have been affected by the deed, as in the 
primary sanctions and to some extent in the secondary sanctions, 
or by removing a conflict within the community itself. The sanc- 
tions are thus of primary significance to sociology in that they 
are reactions on the part of a community to events affecting its 



MANY historical jurists in contrast with the analytical school 
have used the term law to include most if not all processes 
of social control. The term is, however, usually confined 
to 'social control through the systematic application of the force 
of politically organised society' (Pound). The limited application, 
more convenient for purposes of sociological analysis and classi- 
fication, will be adopted in this article; the field of law will there- 
fore be regarded as coterminous with that of organised legal 
sanctions. The obligations imposed on individuals in societies 
where there are no legal sanctions will be regarded as matters of 
custom and convention but not of law; in this sense some simple 
societies have no law, although all have customs which are sup- 
ported by sanctions. 

The confusion which has resulted in the attempt to apply to 
preliterate societies the modern distinction between criminal law 
and civil law can be avoided by making instead the distinction 
between the law of public delicts and the law of private delicts. 
In any society a deed is a public delict if its occurrence normally 
leads to an organised and regular procedure by the whole com- 
munity or by the constituted representatives of social authority, 
which results in the fixing of responsibility upon some person 
within the community and the infliction by the community or by 
its representatives of some hurt or punishment upon the respon- 
sible person. This procedure, which may be called the penal 
sanction, is in its basic form a reaction by the community against 
an action of one of its own members which off'ends some strong 
and definite moral sentiment and thus produces a condition of 
social dysphoria. The immediate function of the reaction is to 
give expression to a collective feeling of moral indignation and so 
to restore the social euphoria. Its ultimate function is to maintain 
the moral sentiment in question at the requisite degree of strength 
in the individuals who constitute the community. 

^ Reprinted from the Encyclopcedia of the Social Sciences, Macmillan Co., 
New York, 1933, Vol. IX, pp. 202-6. 



Comparatively little precise information is available concerning 
penal sanctions in preliterate societies. Among the actions which 
are known to be treated as public delicts in the simpler societies 
are incest, i.e. marriage or sexual congress with persons with whom 
such relations are forbidden; sorcery, or evil magic, by one person 
against another within the community; repeated breaches of 
tribal custom; and various forms of sacrilege. In many preliterate 
societies the penal sanction is applied principally if not solely to 
actions which infringe upon customs regarded by the community 
as sacred, so that the sanction itself may almost be regarded as a 
special form of ritual sanction. Ritual sanctions are derived from 
the belief that certain actions or events render an individual or a 
group ritually unclean, or polluted, so that some specific action 
is required to remove the pollution. In many examples of penal 
sanction it may plausibly be held that a deed such as incest 
produces a pollution of the whole community within which it 
occurs and that the punishment, which may mean the killing of the 
guilty persons, is a means of cleansing the community. Upon the 
establishment of a political or executive authority even of the 
simplest kind disobedience of that authority's commands may be 
subject to penal sanctions and treated as a public delict; moreover, 
direct offences against the constituted authority or against the 
persons in whom that authority rests may be subject to penal 
sanctions. Thus when the social authority rests in chiefs, an 
offence which would be a private delict if committed against a 
commoner may be treated as a public delict when committed 
against a chief. 

In the procedure of a law of private delicts a person or a body 
of persons that has suffered some injury, loss or damage by in- 
fringement of recognised rights appeals to a constituted judicial 
authority, who declares some other person or body of persons 
within the community to be responsible and rules that the de- 
fendant shall give satisfaction to the plaintiff, such satisfaction 
frequently taking the form of the payment of an indemnity or 
damages. A private delict is thus ai action which is subject to 
what may be called a restitutive sanction. The law of private 
delicts in preliterate societies corresponds to the civil law of 
modern times. There are, however, certain important differences. 
In general in modern law actions which fall simply under civil law 
are those which cause damage but are not subject to reprobation. 


Consequently, although the civil sanction expressed through 
the payment of damages causes loss to the defendant, it is not 
specifically punitive. Even in modern civil law, however, a 
magistrate may in special instances award 'punitive damages', 
thereby expressing the view that the injury committed is of such 
a kind as to be properly subject to reprobation and therefore to 
punishment. In modern law when a deed is an offence against 
morality and at the same time inflicts injury it may become 
actionable under both criminal and civil law. The emphasis in the 
punishment for homicide or theft is on its aspect as an offence 
against the community rather than on the principle that resti- 
tution should be made to those who have suffered by the deed. 

In preliterate societies private delicts are for the most part 
killing, wounding, theft, adultery and failure to pay debts; and 
while they are primarily regarded as constituting an injury to 
some member of the community they are subject also to moral 
reprobation as anti-social actions. The sanction is frequently both 
restitutive and repressive, giving satisfaction to the injured person 
and inflicting punishment upon the person responsible for the 
injury; for example, in some African tribes a thief is required 
to restore to the person whom he has robbed double the value 
of what he has taken. In its basic form the law of private delicts 
is a procedure for avoiding or relieving the social dysphoria which 
results from conflicts within a community. An offence committed 
against another member or group of the same community, by 
inflicting a sense of injury upon the victim, creates a disturbance 
of the social life which ceases only when satisfaction is rendered 
to the injured person or persons. Thus in African native law a 
judge is not regarded as having properly settled a case until all 
parties concerned are satisfied with the settlement. 

The distinction between public delicts and private delicts 
illustrates the fact that the law has no single origin. A deed 
committed by a member of the community which offends the 
moral sense of the community may be subject to three sanctions, 
the general or diffuse moral sanction, which makes the guilty 
person subject to the reprobation of his fellows; the ritual sanction, 
which produces in the guilty person a condition of ritual un- 
cleanness that constitutes a danger to himself and to those with 
whom he is in contact — in such cases custom may require him to 
undergo ritual purification or expiation or it may be believed that 


as a result of his sin he will fall ill and die; the penal sanction, 
whereby the community through certain persons acting as its 
constituted judicial authorities inflicts punishment on the guilty 
person, which may be regarded either as a collective expression 
of the moral indignation aroused by the deed or as a means of 
removing the ritual pollution resulting from the deed by imposing 
an expiation upon the guilty person, or as both. 

On the other hand, an action which constitutes an infringe- 
ment of the rights of a person or group of persons may lead to 
retaliation on the part of the injured against the person or group 
responsible for the injury. When such acts of retaliation are 
recognised by custom as justifiable and are subject to a customary 
regulation of procedure, various forms of retaliatory sanctions 
-^ may be said to prevail. In preliterate society generally warfare 
has such a sanction; the waging of war is in some communities, 
as among the Australian hordes, normally an act of retaliation 
carried out by one group against another that is held responsible 
for an injury suffered, and the procedure is regulated by a 
recognised body of customs which is equivalent to the international 
law of modern nations. The institution of organised and regulated 
vengeance is another example of a retaliatory sanction. The 
killing of a man, whether intentional or accidental, constitutes an 
injury to his clan, local community or kindred, for which satis- 
faction is required. The injured group is regarded as justified 
in seeking vengeance and there is frequently an obligation on the 
members of the group to avenge the death. The retaliator}^ action 
is regulated by custom; the lex talionis requires that the damage 
inflicted shall be equivalent to the damage suffered and the prin- 
ciple of collective solidarity permits the avengers to kill a person 
other than the actual murderer, for example his brother, or in some 
instances any member of his clan. When the institution is com- 
pletely organised, custom requires the group responsible for the 
first death to accept the killing of one of their number as an act 
of justice and to make no further retaliation. Retaliatory sanctions 
may also appear in relation to injuries of one person by another; 
for example, the recognised right in certain circumstances of one 
person to challenge another to fight a duel. Among Australian 
tribes an individual who has suffered injury from another may 
by agreement of the elders be given the right to obtain satis- 
faction by throwing spears or boomerangs at him or by spearing 


him in a non- vital part of the body, such as the thigh. In all 
instances of retaliatory sanction there is a customary procedure 
for satisfying the injured person or group whereby resentment 
may be expressed, frequently by inflicting hurt upon the person 
or group responsible for the injury. Where it works effectively 
the result is to provide an expiation for the offence and to remove 
the feeling of injury or resentment in the injured person or persons. 
In many societies retaliation is replaced more or less by a system 
of indemnities; persons or groups having injured other persons or 
groups provide satisfaction to the latter by handing over certain 
valuables. The procedure of providing satisfaction by indemnity 
is widespread in preliterate societies which have not yet developed 
a legal system in the narrow sense. 

Among the Yurok, who are food gatherers and hunters living 
in northern California in small villages with no political organisa- 
tion, there is no regular procedure for dealing with offences 
against the community and therefore no law of public delicts. 
Injuries and offences of one person against another are subject to 
indemnities regulated by custom; every invasion of privilege or 
property must be exactly compensated; for the killing of an in- 
dividual an indemnity or blood money must be paid to the near 
kin. After a feud or war each side must pay for those who have 
been killed on the other side. Only the fact and amount of damage 
are considered; never the question of intent, malice, negligence 
or accident. Once an indemnity for an injury has been accepted 
it is improper for the injured person to harbour any further 
resentment. As the payment of indemnities is arranged by 
negotiation between the persons concerned and not by appeal to 
any judicial authority, the law of private delicts in the strict 
sense is not present. Among the Ifugao, who cultivate rice on 
terraced hillsides of northern Luzon in the Philippines and who 
have no political organisation and no system of clans, 'society does 
not punish injuries to itself except as the censure of public 
opinion is a punishment'; that is, there is no law of public delicts, 
no actual penal sanction. Nevertheless, a person who practises 
sorcery against one of his own kin is put to death by his kin; 
on the other hand, incest between brother and sister, parricide 
and fratricide are said to go unpunished. It is probable, however, 
that there are powerful and effective ritual sanctions against these 
acts. An offence committed by one person against another person 


or an infringement of the rights of one person by another is the 
occasion of a conflict between the kindred of the two parties, 
including relatives through btoh father and mother to the third 
or fourth degree. Retaliation by the killing of the offender or 
sometimes of one of his kin is the regular method of obtaining 
satisfaction in cases of murder, sorcery, adultery discovered in 
flagrante, refusal to pay an indemnity assessed for injury suffered, 
and persistent and wilful refusal to pay a debt when there is ability 
to pay. Satisfaction is provided in other cases by the payment of 
indemnities. There are no judicial authorities before whom 
disputes may be brought; the negotiations are carried out by a 
go-between who belongs to neither of the two opposed groups of 
kindred. Certain persons obtain renown for themselves as success- 
ful go-betweens, but such persons have no authority and are not 
in any sense representatives of the community as a whole. During 
the controversy the two parties are in a condition of ritual enmity 
or opposition and when a settlement is reached they join in a 
peacemaking ceremony. A scale of settlement is recognised by 
custom and in certain circumstances the payments vary according 
to the class — ^wealthy, middle class or poor — to which the group 
receiving or making payment belongs. The Ifugao thus have an 
organised system of justice, which, however, does not constitute 
a system of law in the narrow sense of the term since there is no 
judicial authority. 

An important step is taken towards the formation of a legal 
system where there are recognised arbitrators or judges who hear 
evidence, decide upon responsibility and assess damages; only 
the existence of some authority with power to enforce the judg- 
ments delivered by the judges is then lacking. It has been argued 
plausibly that in some societies a legal system for dealing with 
private delicts has grown up in this manner; disputes are brought 
before arbitrators who declare the custom and apply it to the 
case before them; such courts of arbitration become established 
as regular tribunals; and finally there is developed in the society 
some procedure for enforcing judgments. 

A development similar to this is illustrated by the practices 
of the A-Kamba, A-Kiku}T.i and A-Theraka, Bantu peoples to the 
south and south-east of Mount Kenya in East Africa who live 
in scattered household communities, keep cattle, sheep and goats 
and grow grain in hand-tilled fields. They have no chiefs and are 


divided into well-defined age grades, one of which consists of 
elders who exercise both priestly and judicial functions. If there is 
a dispute in which one person believes his rights have been in- 
fringed by another, the disputants call together a number of elders 
of the district or districts in which they live and these constitute 
a court to hear the case. The court acts primarily as a court of 
arbitration and as a means of deciding upon the customary 
principles of justice by which the dispute should be settled; it 
usually takes no steps to enforce the judgment on the losing party 
but leaves this task to the claimant. In serious cases, however, 
when an offence affects the whole community or when the accused 
is regarded as an habitual and dangerous offender so that public 
indignation makes the affair one of public concern, the elders 
can exercise authority to enforce judgment. The usual procedure 
rests on the ritual powers of the elders; they can pronounce a 
curse, which is feared as inevitably bringing down supernatural 
punishment, on a person who refuses to obey a judgment. The 
killing of a member of one clan by a member of another, whether 
intentional or accidental, is treated by the court of elders as a 
private delict and is settled by the payment of an indemnity to the 
relatives of the victim by the killer and his relatives. The elders 
also possess limited powers of dealing with public delicts by a 
procedure known as kingolle, or mwinge. If a person is held 
guilty of witchcraft or is regarded as an habitual offender and 
thus as a public danger, the elders may inflict the punishment of 
death or may destroy the offender's homestead and expel him from 
the district. Before such action may be taken elders from remote 
regions must be called in for consultation and the consent of near 
relatives of the offender must be obtained. 

The Ashanti afford a contrast to the system of the A-Kamba in 
that they have a well organised law of public delicts, which are 
designated by a native term which means 'things hateful to the 
gods'. These include murder, suicide, certain sexual offences 
including incestuous relations with certain relatives by descent 
and by marriage, certain forms of abuse, assault and stealing, the 
invocation of a curse upon a chief, treason, cowardice, witchcraft, 
the violation of recognised tribal tabus and the breaking of a com- 
mand of the central authority issued and qualified with an oath. 
The Ashanti conception of the law is that all such actions are 
offences against the sacred or supernatural powers on which the 


wellbeing of the whole community depends and that unless these 
offences are expiated by the punishment of the guilty persons the 
whole tribe will suffer. The judicial functions belong to the king 
or chief (the occupant of a sacred stool), before whom the offender 
is tried. The punishment for the more serious offences is de- 
capitation, although in certain circumstances the condemned 
man and his relatives may 'buy his head'; that is, pay a redemp- 
tion price by which his life may be saved. The courts of the chiefs 
do not concern themselves with private delicts, which are denoted 
'household cases' and settled by the authority of heads of kinship 
groups or by negotiations. A dispute concerning a private delict 
may be brought before the chief indirectly if one of the parties 
involved swears an oath, which thus makes the dispute a public 

While the A-Kamba elders are concerned mainly with private 
delicts and the Ashanti chiefs with public delicts, there are tribes 
and nations in Africa and elsewhere in which the central authorities 
— the chiefs or the king and his representatives — administer both 
kinds of law, which may always be differentiated by reference to 
procedure. In the law of private delicts a dispute between persons 
or groups of persons is brought before the judicial tribunal for 
settlement; in the law of public delicts the central authority itself 
and on its own initiative takes action against an offender. Modern 
criminal law and civil law are directly derived respectively from 
the law of public delicts and the law of private delicts; but acts 
which are now regarded as characteristically public delicts, such 
as murder and theft, are in many preliterate societies treated as 
private delicts, while the acts which in such societies are moot 
frequently regarded as public delicts are witchcraft, incest and 

In its most elementary developments law is intimately bound 
up with magic and religion; legal sanctions are closely related to 
ritual sanctions, A full understanding of the beginnings of law 
in simpler societies can therefore be reached only by a comparative 
study of whole systems of social sanctions. 



Los Angeles 
This book is DUE on the last date stamped below. 


4\MK ' 05 }W 


L 005 744 063 8